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 = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 158         static private $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 159         static private $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 160
 161         /**
 162          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 163          * - in group 1 if it is LTR
 164          * - in group 2 if it is RTL
 165          * Does not match if there is no strong directionality codepoint.
 166          *
 167          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 168          *
 169          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 170          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 171          */
 172         // @codeCoverageIgnoreStart
 173         // phpcs:ignore Generic.Files.LineLength
 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
 177         /**
 178          * Get a cached or new language object for a given language code
 179          * @param string $code
 180          * @throws MWException
 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                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 192
 193                 // merge the language object in to get it up front in the cache
 194                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 195                 // get rid of the oldest ones in case we have an overflow
 196                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 197
 198                 return $langObj;
 199         }
 200
 201         /**
 202          * Create a language object for a given language code
 203          * @param string $code
 204          * @param bool $fallback Whether we're going through language fallback chain
 205          * @throws MWException
 206          * @return Language
 207          */
 208         protected static function newFromCode( $code, $fallback = false ) {
 209                 if ( !self::isValidCode( $code ) ) {
 210                         throw new MWException( "Invalid language code \"$code\"" );
 211                 }
 212
 213                 if ( !self::isValidBuiltInCode( $code ) ) {
 214                         // It's not possible to customise this code with class files, so
 215                         // just return a Language object. This is to support uselang= hacks.
 216                         $lang = new Language;
 217                         $lang->setCode( $code );
 218                         return $lang;
 219                 }
 220
 221                 // Check if there is a language class for the code
 222                 $class = self::classFromCode( $code, $fallback );
 223                 if ( class_exists( $class ) ) {
 224                         $lang = new $class;
 225                         return $lang;
 226                 }
 227
 228                 // Keep trying the fallback list until we find an existing class
 229                 $fallbacks = self::getFallbacksFor( $code );
 230                 foreach ( $fallbacks as $fallbackCode ) {
 231                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 232                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 233                         }
 234
 235                         $class = self::classFromCode( $fallbackCode );
 236                         if ( class_exists( $class ) ) {
 237                                 $lang = new $class;
 238                                 $lang->setCode( $code );
 239                                 return $lang;
 240                         }
 241                 }
 242
 243                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 244         }
 245
 246         /**
 247          * Checks whether any localisation is available for that language tag
 248          * in MediaWiki (MessagesXx.php exists).
 249          *
 250          * @param string $code Language tag (in lower case)
 251          * @return bool Whether language is supported
 252          * @since 1.21
 253          */
 254         public static function isSupportedLanguage( $code ) {
 255                 if ( !self::isValidBuiltInCode( $code ) ) {
 256                         return false;
 257                 }
 258
 259                 if ( $code === 'qqq' ) {
 260                         return false;
 261                 }
 262
 263                 return is_readable( self::getMessagesFileName( $code ) ) ||
 264                         is_readable( self::getJsonMessagesFileName( $code ) );
 265         }
 266
 267         /**
 268          * Returns true if a language code string is a well-formed language tag
 269          * according to RFC 5646.
 270          * This function only checks well-formedness; it doesn't check that
 271          * language, script or variant codes actually exist in the repositories.
 272          *
 273          * Based on regexes by Mark Davis of the Unicode Consortium:
 274          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 275          *
 276          * @param string $code
 277          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 278          *
 279          * @return bool
 280          * @since 1.21
 281          */
 282         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 283                 $alpha = '[a-z]';
 284                 $digit = '[0-9]';
 285                 $alphanum = '[a-z0-9]';
 286                 $x = 'x'; # private use singleton
 287                 $singleton = '[a-wy-z]'; # other singleton
 288                 $s = $lenient ? '[-_]' : '-';
 289
 290                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 291                 $script = "$alpha{4}"; # ISO 15924
 292                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 293                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 294                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 295                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 296
 297                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 298                 # Since these are limited, this is safe even later changes to the registry --
 299                 # the only oddity is that it might change the type of the tag, and thus
 300                 # the results from the capturing groups.
 301                 # https://www.iana.org/assignments/language-subtag-registry
 302
 303                 $grandfathered = "en{$s}GB{$s}oed"
 304                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 305                         . "|no{$s}(?:bok|nyn)"
 306                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 307                         . "|zh{$s}min{$s}nan";
 308
 309                 $variantList = "$variant(?:$s$variant)*";
 310                 $extensionList = "$extension(?:$s$extension)*";
 311
 312                 $langtag = "(?:($language)"
 313                         . "(?:$s$script)?"
 314                         . "(?:$s$region)?"
 315                         . "(?:$s$variantList)?"
 316                         . "(?:$s$extensionList)?"
 317                         . "(?:$s$privateUse)?)";
 318
 319                 # The final breakdown, with capturing groups for each of these components
 320                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 321
 322                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 323
 324                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 325         }
 326
 327         /**
 328          * Returns true if a language code string is of a valid form, whether or
 329          * not it exists. This includes codes which are used solely for
 330          * customisation via the MediaWiki namespace.
 331          *
 332          * @param string $code
 333          *
 334          * @return bool
 335          */
 336         public static function isValidCode( $code ) {
 337                 static $cache = [];
 338                 if ( !isset( $cache[$code] ) ) {
 339                         // People think language codes are html safe, so enforce it.
 340                         // Ideally we should only allow a-zA-Z0-9-
 341                         // but, .+ and other chars are often used for {{int:}} hacks
 342                         // see bugs T39564, T39587, T38938
 343                         $cache[$code] =
 344                                 // Protect against path traversal
 345                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 346                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 347                 }
 348                 return $cache[$code];
 349         }
 350
 351         /**
 352          * Returns true if a language code is of a valid form for the purposes of
 353          * internal customisation of MediaWiki, via Messages*.php or *.json.
 354          *
 355          * @param string $code
 356          *
 357          * @throws MWException
 358          * @since 1.18
 359          * @return bool
 360          */
 361         public static function isValidBuiltInCode( $code ) {
 362                 if ( !is_string( $code ) ) {
 363                         if ( is_object( $code ) ) {
 364                                 $addmsg = " of class " . get_class( $code );
 365                         } else {
 366                                 $addmsg = '';
 367                         }
 368                         $type = gettype( $code );
 369                         throw new MWException( __METHOD__ . " must be passed a string, $type given$addmsg" );
 370                 }
 371
 372                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 373         }
 374
 375         /**
 376          * Returns true if a language code is an IETF tag known to MediaWiki.
 377          *
 378          * @param string $tag
 379          *
 380          * @since 1.21
 381          * @return bool
 382          */
 383         public static function isKnownLanguageTag( $tag ) {
 384                 // Quick escape for invalid input to avoid exceptions down the line
 385                 // when code tries to process tags which are not valid at all.
 386                 if ( !self::isValidBuiltInCode( $tag ) ) {
 387                         return false;
 388                 }
 389
 390                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 391                         || self::fetchLanguageName( $tag, $tag ) !== ''
 392                 ) {
 393                         return true;
 394                 }
 395
 396                 return false;
 397         }
 398
 399         /**
 400          * Get the LocalisationCache instance
 401          *
 402          * @return LocalisationCache
 403          */
 404         public static function getLocalisationCache() {
 405                 if ( is_null( self::$dataCache ) ) {
 406                         global $wgLocalisationCacheConf;
 407                         $class = $wgLocalisationCacheConf['class'];
 408                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 409                 }
 410                 return self::$dataCache;
 411         }
 412
 413         function __construct() {
 414                 $this->mConverter = new FakeConverter( $this );
 415                 // Set the code to the name of the descendant
 416                 if ( static::class === 'Language' ) {
 417                         $this->mCode = 'en';
 418                 } else {
 419                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 420                 }
 421                 self::getLocalisationCache();
 422         }
 423
 424         /**
 425          * Reduce memory usage
 426          */
 427         function __destruct() {
 428                 foreach ( $this as $name => $value ) {
 429                         unset( $this->$name );
 430                 }
 431         }
 432
 433         /**
 434          * Hook which will be called if this is the content language.
 435          * Descendants can use this to register hook functions or modify globals
 436          */
 437         function initContLang() {
 438         }
 439
 440         /**
 441          * @return array
 442          * @since 1.19
 443          */
 444         public function getFallbackLanguages() {
 445                 return self::getFallbacksFor( $this->mCode );
 446         }
 447
 448         /**
 449          * Exports $wgBookstoreListEn
 450          * @return array
 451          */
 452         public function getBookstoreList() {
 453                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 454         }
 455
 456         /**
 457          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 458          * available in localised form, it will be included in English.
 459          *
 460          * @return array
 461          */
 462         public function getNamespaces() {
 463                 if ( is_null( $this->namespaceNames ) ) {
 464                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 465
 466                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 467
 468                         $this->namespaceNames = $wgExtraNamespaces +
 469                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 470                         $this->namespaceNames += $validNamespaces;
 471
 472                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 473                         if ( $wgMetaNamespaceTalk ) {
 474                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 475                         } else {
 476                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 477                                 $this->namespaceNames[NS_PROJECT_TALK] =
 478                                         $this->fixVariableInNamespace( $talk );
 479                         }
 480
 481                         # Sometimes a language will be localised but not actually exist on this wiki.
 482                         foreach ( $this->namespaceNames as $key => $text ) {
 483                                 if ( !isset( $validNamespaces[$key] ) ) {
 484                                         unset( $this->namespaceNames[$key] );
 485                                 }
 486                         }
 487
 488                         # The above mixing may leave namespaces out of canonical order.
 489                         # Re-order by namespace ID number...
 490                         ksort( $this->namespaceNames );
 491
 492                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 493                 }
 494
 495                 return $this->namespaceNames;
 496         }
 497
 498         /**
 499          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 500          * @param array $namespaces Array of namespaces (id => name)
 501          */
 502         public function setNamespaces( array $namespaces ) {
 503                 $this->namespaceNames = $namespaces;
 504                 $this->mNamespaceIds = null;
 505         }
 506
 507         /**
 508          * Resets all of the namespace caches. Mainly used for testing
 509          */
 510         public function resetNamespaces() {
 511                 $this->namespaceNames = null;
 512                 $this->mNamespaceIds = null;
 513                 $this->namespaceAliases = null;
 514         }
 515
 516         /**
 517          * A convenience function that returns getNamespaces() with spaces instead of underscores
 518          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 519          *
 520          * @return array
 521          */
 522         public function getFormattedNamespaces() {
 523                 $ns = $this->getNamespaces();
 524                 foreach ( $ns as $k => $v ) {
 525                         $ns[$k] = strtr( $v, '_', ' ' );
 526                 }
 527                 return $ns;
 528         }
 529
 530         /**
 531          * Get a namespace value by key
 532          *
 533          * <code>
 534          * $mw_ns = $wgContLang->getNsText( NS_MEDIAWIKI );
 535          * echo $mw_ns; // prints 'MediaWiki'
 536          * </code>
 537          *
 538          * @param int $index The array key of the namespace to return
 539          * @return string|bool String if the namespace value exists, otherwise false
 540          */
 541         public function getNsText( $index ) {
 542                 $ns = $this->getNamespaces();
 543                 return $ns[$index] ?? false;
 544         }
 545
 546         /**
 547          * A convenience function that returns the same thing as
 548          * getNsText() except with '_' changed to ' ', useful for
 549          * producing output.
 550          *
 551          * <code>
 552          * $mw_ns = $wgContLang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 553          * echo $mw_ns; // prints 'MediaWiki talk'
 554          * </code>
 555          *
 556          * @param int $index The array key of the namespace to return
 557          * @return string Namespace name without underscores (empty string if namespace does not exist)
 558          */
 559         public function getFormattedNsText( $index ) {
 560                 $ns = $this->getNsText( $index );
 561                 return strtr( $ns, '_', ' ' );
 562         }
 563
 564         /**
 565          * Returns gender-dependent namespace alias if available.
 566          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 567          * @param int $index Namespace index
 568          * @param string $gender Gender key (male, female... )
 569          * @return string
 570          * @since 1.18
 571          */
 572         public function getGenderNsText( $index, $gender ) {
 573                 global $wgExtraGenderNamespaces;
 574
 575                 $ns = $wgExtraGenderNamespaces +
 576                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 577
 578                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 579         }
 580
 581         /**
 582          * Whether this language uses gender-dependent namespace aliases.
 583          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 584          * @return bool
 585          * @since 1.18
 586          */
 587         public function needsGenderDistinction() {
 588                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 589                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 590                         // $wgExtraGenderNamespaces overrides everything
 591                         return true;
 592                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 593                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 594                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 595                         return false;
 596                 } else {
 597                         // Check what is in i18n files
 598                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 599                         return count( $aliases ) > 0;
 600                 }
 601         }
 602
 603         /**
 604          * Get a namespace key by value, case insensitive.
 605          * Only matches namespace names for the current language, not the
 606          * canonical ones defined in Namespace.php.
 607          *
 608          * @param string $text
 609          * @return int|bool An integer if $text is a valid value otherwise false
 610          */
 611         function getLocalNsIndex( $text ) {
 612                 $lctext = $this->lc( $text );
 613                 $ids = $this->getNamespaceIds();
 614                 return $ids[$lctext] ?? false;
 615         }
 616
 617         /**
 618          * @return array
 619          */
 620         public function getNamespaceAliases() {
 621                 if ( is_null( $this->namespaceAliases ) ) {
 622                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 623                         if ( !$aliases ) {
 624                                 $aliases = [];
 625                         } else {
 626                                 foreach ( $aliases as $name => $index ) {
 627                                         if ( $index === NS_PROJECT_TALK ) {
 628                                                 unset( $aliases[$name] );
 629                                                 $name = $this->fixVariableInNamespace( $name );
 630                                                 $aliases[$name] = $index;
 631                                         }
 632                                 }
 633                         }
 634
 635                         global $wgExtraGenderNamespaces;
 636                         $genders = $wgExtraGenderNamespaces +
 637                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 638                         foreach ( $genders as $index => $forms ) {
 639                                 foreach ( $forms as $alias ) {
 640                                         $aliases[$alias] = $index;
 641                                 }
 642                         }
 643
 644                         # Also add converted namespace names as aliases, to avoid confusion.
 645                         $convertedNames = [];
 646                         foreach ( $this->getVariants() as $variant ) {
 647                                 if ( $variant === $this->mCode ) {
 648                                         continue;
 649                                 }
 650                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 651                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 652                                 }
 653                         }
 654
 655                         $this->namespaceAliases = $aliases + $convertedNames;
 656                 }
 657
 658                 return $this->namespaceAliases;
 659         }
 660
 661         /**
 662          * @return array
 663          */
 664         public function getNamespaceIds() {
 665                 if ( is_null( $this->mNamespaceIds ) ) {
 666                         global $wgNamespaceAliases;
 667                         # Put namespace names and aliases into a hashtable.
 668                         # If this is too slow, then we should arrange it so that it is done
 669                         # before caching. The catch is that at pre-cache time, the above
 670                         # class-specific fixup hasn't been done.
 671                         $this->mNamespaceIds = [];
 672                         foreach ( $this->getNamespaces() as $index => $name ) {
 673                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 674                         }
 675                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 676                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 677                         }
 678                         if ( $wgNamespaceAliases ) {
 679                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 680                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 681                                 }
 682                         }
 683                 }
 684                 return $this->mNamespaceIds;
 685         }
 686
 687         /**
 688          * Get a namespace key by value, case insensitive.  Canonical namespace
 689          * names override custom ones defined for the current language.
 690          *
 691          * @param string $text
 692          * @return int|bool An integer if $text is a valid value otherwise false
 693          */
 694         public function getNsIndex( $text ) {
 695                 $lctext = $this->lc( $text );
 696                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 697                 if ( $ns !== null ) {
 698                         return $ns;
 699                 }
 700                 $ids = $this->getNamespaceIds();
 701                 return $ids[$lctext] ?? false;
 702         }
 703
 704         /**
 705          * short names for language variants used for language conversion links.
 706          *
 707          * @param string $code
 708          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 709          * @return string
 710          */
 711         public function getVariantname( $code, $usemsg = true ) {
 712                 $msg = "variantname-$code";
 713                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 714                         return $this->getMessageFromDB( $msg );
 715                 }
 716                 $name = self::fetchLanguageName( $code );
 717                 if ( $name ) {
 718                         return $name; # if it's defined as a language name, show that
 719                 } else {
 720                         # otherwise, output the language code
 721                         return $code;
 722                 }
 723         }
 724
 725         /**
 726          * @return string[]|bool List of date format preference keys, or false if disabled.
 727          */
 728         public function getDatePreferences() {
 729                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 730         }
 731
 732         /**
 733          * @return array
 734          */
 735         function getDateFormats() {
 736                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 737         }
 738
 739         /**
 740          * @return array|string
 741          */
 742         public function getDefaultDateFormat() {
 743                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 744                 if ( $df === 'dmy or mdy' ) {
 745                         global $wgAmericanDates;
 746                         return $wgAmericanDates ? 'mdy' : 'dmy';
 747                 } else {
 748                         return $df;
 749                 }
 750         }
 751
 752         /**
 753          * @return array
 754          */
 755         public function getDatePreferenceMigrationMap() {
 756                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 757         }
 758
 759         /**
 760          * @param string $image
 761          * @return array|null
 762          */
 763         function getImageFile( $image ) {
 764                 return self::$dataCache->getSubitem( $this->mCode, 'imageFiles', $image );
 765         }
 766
 767         /**
 768          * @return array
 769          * @since 1.24
 770          */
 771         public function getImageFiles() {
 772                 return self::$dataCache->getItem( $this->mCode, 'imageFiles' );
 773         }
 774
 775         /**
 776          * @return array
 777          */
 778         public function getExtraUserToggles() {
 779                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 780         }
 781
 782         /**
 783          * @param string $tog
 784          * @return string
 785          */
 786         function getUserToggle( $tog ) {
 787                 return $this->getMessageFromDB( "tog-$tog" );
 788         }
 789
 790         /**
 791          * Get an array of language names, indexed by code.
 792          * @param null|string $inLanguage Code of language in which to return the names
 793          *              Use null for autonyms (native names)
 794          * @param string $include One of:
 795          *              'all' all available languages
 796          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 797          *              'mwfile' only if the language is in 'mw' *and* has a message file
 798          * @return array Language code => language name (sorted by key)
 799          * @since 1.20
 800          */
 801         public static function fetchLanguageNames( $inLanguage = null, $include = 'mw' ) {
 802                 $cacheKey = $inLanguage === null ? 'null' : $inLanguage;
 803                 $cacheKey .= ":$include";
 804                 if ( self::$languageNameCache === null ) {
 805                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 806                 }
 807
 808                 $ret = self::$languageNameCache->get( $cacheKey );
 809                 if ( !$ret ) {
 810                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 811                         self::$languageNameCache->set( $cacheKey, $ret );
 812                 }
 813                 return $ret;
 814         }
 815
 816         /**
 817          * Uncached helper for fetchLanguageNames
 818          * @param null|string $inLanguage Code of language in which to return the names
 819          *              Use null for autonyms (native names)
 820          * @param string $include One of:
 821          *              'all' all available languages
 822          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 823          *              'mwfile' only if the language is in 'mw' *and* has a message file
 824          * @return array Language code => language name (sorted by key)
 825          */
 826         private static function fetchLanguageNamesUncached( $inLanguage = null, $include = 'mw' ) {
 827                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 828
 829                 // If passed an invalid language code to use, fallback to en
 830                 if ( $inLanguage !== null && !self::isValidCode( $inLanguage ) ) {
 831                         $inLanguage = 'en';
 832                 }
 833
 834                 $names = [];
 835
 836                 if ( $inLanguage ) {
 837                         # TODO: also include when $inLanguage is null, when this code is more efficient
 838                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 839                 }
 840
 841                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 842                 if ( $wgUsePigLatinVariant ) {
 843                         // Pig Latin (for variant development)
 844                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 845                 }
 846
 847                 foreach ( $mwNames as $mwCode => $mwName ) {
 848                         # - Prefer own MediaWiki native name when not using the hook
 849                         # - For other names just add if not added through the hook
 850                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 851                                 $names[$mwCode] = $mwName;
 852                         }
 853                 }
 854
 855                 if ( $include === 'all' ) {
 856                         ksort( $names );
 857                         return $names;
 858                 }
 859
 860                 $returnMw = [];
 861                 $coreCodes = array_keys( $mwNames );
 862                 foreach ( $coreCodes as $coreCode ) {
 863                         $returnMw[$coreCode] = $names[$coreCode];
 864                 }
 865
 866                 if ( $include === 'mwfile' ) {
 867                         $namesMwFile = [];
 868                         # We do this using a foreach over the codes instead of a directory
 869                         # loop so that messages files in extensions will work correctly.
 870                         foreach ( $returnMw as $code => $value ) {
 871                                 if ( is_readable( self::getMessagesFileName( $code ) )
 872                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 873                                 ) {
 874                                         $namesMwFile[$code] = $names[$code];
 875                                 }
 876                         }
 877
 878                         ksort( $namesMwFile );
 879                         return $namesMwFile;
 880                 }
 881
 882                 ksort( $returnMw );
 883                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 884                 return $returnMw;
 885         }
 886
 887         /**
 888          * @param string $code The code of the language for which to get the name
 889          * @param null|string $inLanguage Code of language in which to return the name (null for autonyms)
 890          * @param string $include 'all', 'mw' or 'mwfile'; see fetchLanguageNames()
 891          * @return string Language name or empty
 892          * @since 1.20
 893          */
 894         public static function fetchLanguageName( $code, $inLanguage = null, $include = 'all' ) {
 895                 $code = strtolower( $code );
 896                 $array = self::fetchLanguageNames( $inLanguage, $include );
 897                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 898         }
 899
 900         /**
 901          * Get a message from the MediaWiki namespace.
 902          *
 903          * @param string $msg Message name
 904          * @return string
 905          */
 906         public function getMessageFromDB( $msg ) {
 907                 return $this->msg( $msg )->text();
 908         }
 909
 910         /**
 911          * Get message object in this language. Only for use inside this class.
 912          *
 913          * @param string $msg Message name
 914          * @return Message
 915          */
 916         protected function msg( $msg ) {
 917                 return wfMessage( $msg )->inLanguage( $this );
 918         }
 919
 920         /**
 921          * @param string $key
 922          * @return string
 923          */
 924         public function getMonthName( $key ) {
 925                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 926         }
 927
 928         /**
 929          * @return array
 930          */
 931         public function getMonthNamesArray() {
 932                 $monthNames = [ '' ];
 933                 for ( $i = 1; $i < 13; $i++ ) {
 934                         $monthNames[] = $this->getMonthName( $i );
 935                 }
 936                 return $monthNames;
 937         }
 938
 939         /**
 940          * @param string $key
 941          * @return string
 942          */
 943         public function getMonthNameGen( $key ) {
 944                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 945         }
 946
 947         /**
 948          * @param string $key
 949          * @return string
 950          */
 951         public function getMonthAbbreviation( $key ) {
 952                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 953         }
 954
 955         /**
 956          * @return array
 957          */
 958         public function getMonthAbbreviationsArray() {
 959                 $monthNames = [ '' ];
 960                 for ( $i = 1; $i < 13; $i++ ) {
 961                         $monthNames[] = $this->getMonthAbbreviation( $i );
 962                 }
 963                 return $monthNames;
 964         }
 965
 966         /**
 967          * @param string $key
 968          * @return string
 969          */
 970         public function getWeekdayName( $key ) {
 971                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
 972         }
 973
 974         /**
 975          * @param string $key
 976          * @return string
 977          */
 978         function getWeekdayAbbreviation( $key ) {
 979                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
 980         }
 981
 982         /**
 983          * @param string $key
 984          * @return string
 985          */
 986         function getIranianCalendarMonthName( $key ) {
 987                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
 988         }
 989
 990         /**
 991          * @param string $key
 992          * @return string
 993          */
 994         function getHebrewCalendarMonthName( $key ) {
 995                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
 996         }
 997
 998         /**
 999          * @param string $key
1000          * @return string
1001          */
1002         function getHebrewCalendarMonthNameGen( $key ) {
1003                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1004         }
1005
1006         /**
1007          * @param string $key
1008          * @return string
1009          */
1010         function getHijriCalendarMonthName( $key ) {
1011                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1012         }
1013
1014         /**
1015          * Pass through result from $dateTimeObj->format()
1016          * @param DateTime|bool|null &$dateTimeObj
1017          * @param string $ts
1018          * @param DateTimeZone|bool|null $zone
1019          * @param string $code
1020          * @return string
1021          */
1022         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1023                 if ( !$dateTimeObj ) {
1024                         $dateTimeObj = DateTime::createFromFormat(
1025                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1026                         );
1027                 }
1028                 return $dateTimeObj->format( $code );
1029         }
1030
1031         /**
1032          * This is a workalike of PHP's date() function, but with better
1033          * internationalisation, a reduced set of format characters, and a better
1034          * escaping format.
1035          *
1036          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1037          * the PHP manual for definitions. There are a number of extensions, which
1038          * start with "x":
1039          *
1040          *    xn   Do not translate digits of the next numeric format character
1041          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1042          *    xr   Use roman numerals for the next numeric format character
1043          *    xh   Use hebrew numerals for the next numeric format character
1044          *    xx   Literal x
1045          *    xg   Genitive month name
1046          *
1047          *    xij  j (day number) in Iranian calendar
1048          *    xiF  F (month name) in Iranian calendar
1049          *    xin  n (month number) in Iranian calendar
1050          *    xiy  y (two digit year) in Iranian calendar
1051          *    xiY  Y (full year) in Iranian calendar
1052          *    xit  t (days in month) in Iranian calendar
1053          *    xiz  z (day of the year) in Iranian calendar
1054          *
1055          *    xjj  j (day number) in Hebrew calendar
1056          *    xjF  F (month name) in Hebrew calendar
1057          *    xjt  t (days in month) in Hebrew calendar
1058          *    xjx  xg (genitive month name) in Hebrew calendar
1059          *    xjn  n (month number) in Hebrew calendar
1060          *    xjY  Y (full year) in Hebrew calendar
1061          *
1062          *    xmj  j (day number) in Hijri calendar
1063          *    xmF  F (month name) in Hijri calendar
1064          *    xmn  n (month number) in Hijri calendar
1065          *    xmY  Y (full year) in Hijri calendar
1066          *
1067          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1068          *                       identical to the Gregorian calendar
1069          *    xoY  Y (full year) in Minguo calendar or Juche year.
1070          *                       Months and days are identical to the
1071          *                       Gregorian calendar
1072          *    xtY  Y (full year) in Japanese nengo. Months and days are
1073          *                       identical to the Gregorian calendar
1074          *
1075          * Characters enclosed in double quotes will be considered literal (with
1076          * the quotes themselves removed). Unmatched quotes will be considered
1077          * literal quotes. Example:
1078          *
1079          * "The month is" F       => The month is January
1080          * i's"                   => 20'11"
1081          *
1082          * Backslash escaping is also supported.
1083          *
1084          * Input timestamp is assumed to be pre-normalized to the desired local
1085          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1086          * $ts is UTC if $zone is not given.
1087          *
1088          * @param string $format
1089          * @param string $ts 14-character timestamp
1090          *      YYYYMMDDHHMMSS
1091          *      01234567890123
1092          * @param DateTimeZone $zone Timezone of $ts
1093          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1094          * Only makes sense if $ts is the current time.
1095          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1096          *
1097          * @throws MWException
1098          * @return string
1099          */
1100         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1101                 $s = '';
1102                 $raw = false;
1103                 $roman = false;
1104                 $hebrewNum = false;
1105                 $dateTimeObj = false;
1106                 $rawToggle = false;
1107                 $iranian = false;
1108                 $hebrew = false;
1109                 $hijri = false;
1110                 $thai = false;
1111                 $minguo = false;
1112                 $tenno = false;
1113
1114                 $usedSecond = false;
1115                 $usedMinute = false;
1116                 $usedHour = false;
1117                 $usedAMPM = false;
1118                 $usedDay = false;
1119                 $usedWeek = false;
1120                 $usedMonth = false;
1121                 $usedYear = false;
1122                 $usedISOYear = false;
1123                 $usedIsLeapYear = false;
1124
1125                 $usedHebrewMonth = false;
1126                 $usedIranianMonth = false;
1127                 $usedHijriMonth = false;
1128                 $usedHebrewYear = false;
1129                 $usedIranianYear = false;
1130                 $usedHijriYear = false;
1131                 $usedTennoYear = false;
1132
1133                 if ( strlen( $ts ) !== 14 ) {
1134                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1135                 }
1136
1137                 if ( !ctype_digit( $ts ) ) {
1138                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1139                 }
1140
1141                 $formatLength = strlen( $format );
1142                 for ( $p = 0; $p < $formatLength; $p++ ) {
1143                         $num = false;
1144                         $code = $format[$p];
1145                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1146                                 $code .= $format[++$p];
1147                         }
1148
1149                         if ( ( $code === 'xi'
1150                                         || $code === 'xj'
1151                                         || $code === 'xk'
1152                                         || $code === 'xm'
1153                                         || $code === 'xo'
1154                                         || $code === 'xt' )
1155                                 && $p < $formatLength - 1 ) {
1156                                 $code .= $format[++$p];
1157                         }
1158
1159                         switch ( $code ) {
1160                                 case 'xx':
1161                                         $s .= 'x';
1162                                         break;
1163                                 case 'xn':
1164                                         $raw = true;
1165                                         break;
1166                                 case 'xN':
1167                                         $rawToggle = !$rawToggle;
1168                                         break;
1169                                 case 'xr':
1170                                         $roman = true;
1171                                         break;
1172                                 case 'xh':
1173                                         $hebrewNum = true;
1174                                         break;
1175                                 case 'xg':
1176                                         $usedMonth = true;
1177                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1178                                         break;
1179                                 case 'xjx':
1180                                         $usedHebrewMonth = true;
1181                                         if ( !$hebrew ) {
1182                                                 $hebrew = self::tsToHebrew( $ts );
1183                                         }
1184                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1185                                         break;
1186                                 case 'd':
1187                                         $usedDay = true;
1188                                         $num = substr( $ts, 6, 2 );
1189                                         break;
1190                                 case 'D':
1191                                         $usedDay = true;
1192                                         $s .= $this->getWeekdayAbbreviation(
1193                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1194                                         );
1195                                         break;
1196                                 case 'j':
1197                                         $usedDay = true;
1198                                         $num = intval( substr( $ts, 6, 2 ) );
1199                                         break;
1200                                 case 'xij':
1201                                         $usedDay = true;
1202                                         if ( !$iranian ) {
1203                                                 $iranian = self::tsToIranian( $ts );
1204                                         }
1205                                         $num = $iranian[2];
1206                                         break;
1207                                 case 'xmj':
1208                                         $usedDay = true;
1209                                         if ( !$hijri ) {
1210                                                 $hijri = self::tsToHijri( $ts );
1211                                         }
1212                                         $num = $hijri[2];
1213                                         break;
1214                                 case 'xjj':
1215                                         $usedDay = true;
1216                                         if ( !$hebrew ) {
1217                                                 $hebrew = self::tsToHebrew( $ts );
1218                                         }
1219                                         $num = $hebrew[2];
1220                                         break;
1221                                 case 'l':
1222                                         $usedDay = true;
1223                                         $s .= $this->getWeekdayName(
1224                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1225                                         );
1226                                         break;
1227                                 case 'F':
1228                                         $usedMonth = true;
1229                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1230                                         break;
1231                                 case 'xiF':
1232                                         $usedIranianMonth = true;
1233                                         if ( !$iranian ) {
1234                                                 $iranian = self::tsToIranian( $ts );
1235                                         }
1236                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1237                                         break;
1238                                 case 'xmF':
1239                                         $usedHijriMonth = true;
1240                                         if ( !$hijri ) {
1241                                                 $hijri = self::tsToHijri( $ts );
1242                                         }
1243                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1244                                         break;
1245                                 case 'xjF':
1246                                         $usedHebrewMonth = true;
1247                                         if ( !$hebrew ) {
1248                                                 $hebrew = self::tsToHebrew( $ts );
1249                                         }
1250                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1251                                         break;
1252                                 case 'm':
1253                                         $usedMonth = true;
1254                                         $num = substr( $ts, 4, 2 );
1255                                         break;
1256                                 case 'M':
1257                                         $usedMonth = true;
1258                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1259                                         break;
1260                                 case 'n':
1261                                         $usedMonth = true;
1262                                         $num = intval( substr( $ts, 4, 2 ) );
1263                                         break;
1264                                 case 'xin':
1265                                         $usedIranianMonth = true;
1266                                         if ( !$iranian ) {
1267                                                 $iranian = self::tsToIranian( $ts );
1268                                         }
1269                                         $num = $iranian[1];
1270                                         break;
1271                                 case 'xmn':
1272                                         $usedHijriMonth = true;
1273                                         if ( !$hijri ) {
1274                                                 $hijri = self::tsToHijri( $ts );
1275                                         }
1276                                         $num = $hijri[1];
1277                                         break;
1278                                 case 'xjn':
1279                                         $usedHebrewMonth = true;
1280                                         if ( !$hebrew ) {
1281                                                 $hebrew = self::tsToHebrew( $ts );
1282                                         }
1283                                         $num = $hebrew[1];
1284                                         break;
1285                                 case 'xjt':
1286                                         $usedHebrewMonth = true;
1287                                         if ( !$hebrew ) {
1288                                                 $hebrew = self::tsToHebrew( $ts );
1289                                         }
1290                                         $num = $hebrew[3];
1291                                         break;
1292                                 case 'Y':
1293                                         $usedYear = true;
1294                                         $num = substr( $ts, 0, 4 );
1295                                         break;
1296                                 case 'xiY':
1297                                         $usedIranianYear = true;
1298                                         if ( !$iranian ) {
1299                                                 $iranian = self::tsToIranian( $ts );
1300                                         }
1301                                         $num = $iranian[0];
1302                                         break;
1303                                 case 'xmY':
1304                                         $usedHijriYear = true;
1305                                         if ( !$hijri ) {
1306                                                 $hijri = self::tsToHijri( $ts );
1307                                         }
1308                                         $num = $hijri[0];
1309                                         break;
1310                                 case 'xjY':
1311                                         $usedHebrewYear = true;
1312                                         if ( !$hebrew ) {
1313                                                 $hebrew = self::tsToHebrew( $ts );
1314                                         }
1315                                         $num = $hebrew[0];
1316                                         break;
1317                                 case 'xkY':
1318                                         $usedYear = true;
1319                                         if ( !$thai ) {
1320                                                 $thai = self::tsToYear( $ts, 'thai' );
1321                                         }
1322                                         $num = $thai[0];
1323                                         break;
1324                                 case 'xoY':
1325                                         $usedYear = true;
1326                                         if ( !$minguo ) {
1327                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1328                                         }
1329                                         $num = $minguo[0];
1330                                         break;
1331                                 case 'xtY':
1332                                         $usedTennoYear = true;
1333                                         if ( !$tenno ) {
1334                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1335                                         }
1336                                         $num = $tenno[0];
1337                                         break;
1338                                 case 'y':
1339                                         $usedYear = true;
1340                                         $num = substr( $ts, 2, 2 );
1341                                         break;
1342                                 case 'xiy':
1343                                         $usedIranianYear = true;
1344                                         if ( !$iranian ) {
1345                                                 $iranian = self::tsToIranian( $ts );
1346                                         }
1347                                         $num = substr( $iranian[0], -2 );
1348                                         break;
1349                                 case 'xit':
1350                                         $usedIranianYear = true;
1351                                         if ( !$iranian ) {
1352                                                 $iranian = self::tsToIranian( $ts );
1353                                         }
1354                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1355                                         break;
1356                                 case 'xiz':
1357                                         $usedIranianYear = true;
1358                                         if ( !$iranian ) {
1359                                                 $iranian = self::tsToIranian( $ts );
1360                                         }
1361                                         $num = $iranian[3];
1362                                         break;
1363                                 case 'a':
1364                                         $usedAMPM = true;
1365                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1366                                         break;
1367                                 case 'A':
1368                                         $usedAMPM = true;
1369                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1370                                         break;
1371                                 case 'g':
1372                                         $usedHour = true;
1373                                         $h = substr( $ts, 8, 2 );
1374                                         $num = $h % 12 ? $h % 12 : 12;
1375                                         break;
1376                                 case 'G':
1377                                         $usedHour = true;
1378                                         $num = intval( substr( $ts, 8, 2 ) );
1379                                         break;
1380                                 case 'h':
1381                                         $usedHour = true;
1382                                         $h = substr( $ts, 8, 2 );
1383                                         $num = sprintf( '%02d', $h % 12 ? $h % 12 : 12 );
1384                                         break;
1385                                 case 'H':
1386                                         $usedHour = true;
1387                                         $num = substr( $ts, 8, 2 );
1388                                         break;
1389                                 case 'i':
1390                                         $usedMinute = true;
1391                                         $num = substr( $ts, 10, 2 );
1392                                         break;
1393                                 case 's':
1394                                         $usedSecond = true;
1395                                         $num = substr( $ts, 12, 2 );
1396                                         break;
1397                                 case 'c':
1398                                 case 'r':
1399                                         $usedSecond = true;
1400                                         // fall through
1401                                 case 'e':
1402                                 case 'O':
1403                                 case 'P':
1404                                 case 'T':
1405                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1406                                         break;
1407                                 case 'w':
1408                                 case 'N':
1409                                 case 'z':
1410                                         $usedDay = true;
1411                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1412                                         break;
1413                                 case 'W':
1414                                         $usedWeek = true;
1415                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1416                                         break;
1417                                 case 't':
1418                                         $usedMonth = true;
1419                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1420                                         break;
1421                                 case 'L':
1422                                         $usedIsLeapYear = true;
1423                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1424                                         break;
1425                                 case 'o':
1426                                         $usedISOYear = true;
1427                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1428                                         break;
1429                                 case 'U':
1430                                         $usedSecond = true;
1431                                         // fall through
1432                                 case 'I':
1433                                 case 'Z':
1434                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1435                                         break;
1436                                 case '\\':
1437                                         # Backslash escaping
1438                                         if ( $p < $formatLength - 1 ) {
1439                                                 $s .= $format[++$p];
1440                                         } else {
1441                                                 $s .= '\\';
1442                                         }
1443                                         break;
1444                                 case '"':
1445                                         # Quoted literal
1446                                         if ( $p < $formatLength - 1 ) {
1447                                                 $endQuote = strpos( $format, '"', $p + 1 );
1448                                                 if ( $endQuote === false ) {
1449                                                         # No terminating quote, assume literal "
1450                                                         $s .= '"';
1451                                                 } else {
1452                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1453                                                         $p = $endQuote;
1454                                                 }
1455                                         } else {
1456                                                 # Quote at end of string, assume literal "
1457                                                 $s .= '"';
1458                                         }
1459                                         break;
1460                                 default:
1461                                         $s .= $format[$p];
1462                         }
1463                         if ( $num !== false ) {
1464                                 if ( $rawToggle || $raw ) {
1465                                         $s .= $num;
1466                                         $raw = false;
1467                                 } elseif ( $roman ) {
1468                                         $s .= self::romanNumeral( $num );
1469                                         $roman = false;
1470                                 } elseif ( $hebrewNum ) {
1471                                         $s .= self::hebrewNumeral( $num );
1472                                         $hebrewNum = false;
1473                                 } else {
1474                                         $s .= $this->formatNum( $num, true );
1475                                 }
1476                         }
1477                 }
1478
1479                 if ( $ttl === 'unused' ) {
1480                         // No need to calculate the TTL, the caller wont use it anyway.
1481                 } elseif ( $usedSecond ) {
1482                         $ttl = 1;
1483                 } elseif ( $usedMinute ) {
1484                         $ttl = 60 - substr( $ts, 12, 2 );
1485                 } elseif ( $usedHour ) {
1486                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1487                 } elseif ( $usedAMPM ) {
1488                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1489                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1490                 } elseif (
1491                         $usedDay ||
1492                         $usedHebrewMonth ||
1493                         $usedIranianMonth ||
1494                         $usedHijriMonth ||
1495                         $usedHebrewYear ||
1496                         $usedIranianYear ||
1497                         $usedHijriYear ||
1498                         $usedTennoYear
1499                 ) {
1500                         // @todo Someone who understands the non-Gregorian calendars
1501                         // should write proper logic for them so that they don't need purged every day.
1502                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1503                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1504                 } else {
1505                         $possibleTtls = [];
1506                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1507                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1508                         if ( $usedWeek ) {
1509                                 $possibleTtls[] =
1510                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1511                                         $timeRemainingInDay;
1512                         } elseif ( $usedISOYear ) {
1513                                 // December 28th falls on the last ISO week of the year, every year.
1514                                 // The last ISO week of a year can be 52 or 53.
1515                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1516                                         'Ymd',
1517                                         substr( $ts, 0, 4 ) . '1228',
1518                                         $zone ?: new DateTimeZone( 'UTC' )
1519                                 )->format( 'W' );
1520                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1521                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1522                                 $timeRemainingInWeek =
1523                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1524                                         + $timeRemainingInDay;
1525                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1526                         }
1527
1528                         if ( $usedMonth ) {
1529                                 $possibleTtls[] =
1530                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1531                                                 substr( $ts, 6, 2 ) ) * 86400
1532                                         + $timeRemainingInDay;
1533                         } elseif ( $usedYear ) {
1534                                 $possibleTtls[] =
1535                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1536                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1537                                         + $timeRemainingInDay;
1538                         } elseif ( $usedIsLeapYear ) {
1539                                 $year = substr( $ts, 0, 4 );
1540                                 $timeRemainingInYear =
1541                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1542                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1543                                         + $timeRemainingInDay;
1544                                 $mod = $year % 4;
1545                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1546                                         // this isn't a leap year. see when the next one starts
1547                                         $nextCandidate = $year - $mod + 4;
1548                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1549                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1550                                                         $timeRemainingInYear;
1551                                         } else {
1552                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1553                                                         $timeRemainingInYear;
1554                                         }
1555                                 } else {
1556                                         // this is a leap year, so the next year isn't
1557                                         $possibleTtls[] = $timeRemainingInYear;
1558                                 }
1559                         }
1560
1561                         if ( $possibleTtls ) {
1562                                 $ttl = min( $possibleTtls );
1563                         }
1564                 }
1565
1566                 return $s;
1567         }
1568
1569         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1570         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1571
1572         /**
1573          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1574          * Gregorian dates to Iranian dates. Originally written in C, it
1575          * is released under the terms of GNU Lesser General Public
1576          * License. Conversion to PHP was performed by Niklas Laxström.
1577          *
1578          * Link: http://www.farsiweb.info/jalali/jalali.c
1579          *
1580          * @param string $ts
1581          *
1582          * @return int[]
1583          */
1584         private static function tsToIranian( $ts ) {
1585                 $gy = substr( $ts, 0, 4 ) - 1600;
1586                 $gm = substr( $ts, 4, 2 ) - 1;
1587                 $gd = substr( $ts, 6, 2 ) - 1;
1588
1589                 # Days passed from the beginning (including leap years)
1590                 $gDayNo = 365 * $gy
1591                         + floor( ( $gy + 3 ) / 4 )
1592                         - floor( ( $gy + 99 ) / 100 )
1593                         + floor( ( $gy + 399 ) / 400 );
1594
1595                 // Add days of the past months of this year
1596                 for ( $i = 0; $i < $gm; $i++ ) {
1597                         $gDayNo += self::$GREG_DAYS[$i];
1598                 }
1599
1600                 // Leap years
1601                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1602                         $gDayNo++;
1603                 }
1604
1605                 // Days passed in current month
1606                 $gDayNo += (int)$gd;
1607
1608                 $jDayNo = $gDayNo - 79;
1609
1610                 $jNp = floor( $jDayNo / 12053 );
1611                 $jDayNo %= 12053;
1612
1613                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1614                 $jDayNo %= 1461;
1615
1616                 if ( $jDayNo >= 366 ) {
1617                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1618                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1619                 }
1620
1621                 $jz = $jDayNo;
1622
1623                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1624                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1625                 }
1626
1627                 $jm = $i + 1;
1628                 $jd = $jDayNo + 1;
1629
1630                 return [ $jy, $jm, $jd, $jz ];
1631         }
1632
1633         /**
1634          * Converting Gregorian dates to Hijri dates.
1635          *
1636          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1637          *
1638          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1639          *
1640          * @param string $ts
1641          *
1642          * @return int[]
1643          */
1644         private static function tsToHijri( $ts ) {
1645                 $year = substr( $ts, 0, 4 );
1646                 $month = substr( $ts, 4, 2 );
1647                 $day = substr( $ts, 6, 2 );
1648
1649                 $zyr = $year;
1650                 $zd = $day;
1651                 $zm = $month;
1652                 $zy = $zyr;
1653
1654                 if (
1655                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1656                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1657                 ) {
1658                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1659                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1660                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1661                                         $zd - 32075;
1662                 } else {
1663                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1664                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1665                 }
1666
1667                 $zl = $zjd - 1948440 + 10632;
1668                 $zn = (int)( ( $zl - 1 ) / 10631 );
1669                 $zl = $zl - 10631 * $zn + 354;
1670                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1671                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1672                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1673                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1674                 $zm = (int)( ( 24 * $zl ) / 709 );
1675                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1676                 $zy = 30 * $zn + $zj - 30;
1677
1678                 return [ $zy, $zm, $zd ];
1679         }
1680
1681         /**
1682          * Converting Gregorian dates to Hebrew dates.
1683          *
1684          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1685          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1686          * to translate the relevant functions into PHP and release them under
1687          * GNU GPL.
1688          *
1689          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1690          * and Adar II is 14. In a non-leap year, Adar is 6.
1691          *
1692          * @param string $ts
1693          *
1694          * @return int[]
1695          */
1696         private static function tsToHebrew( $ts ) {
1697                 # Parse date
1698                 $year = substr( $ts, 0, 4 );
1699                 $month = substr( $ts, 4, 2 );
1700                 $day = substr( $ts, 6, 2 );
1701
1702                 # Calculate Hebrew year
1703                 $hebrewYear = $year + 3760;
1704
1705                 # Month number when September = 1, August = 12
1706                 $month += 4;
1707                 if ( $month > 12 ) {
1708                         # Next year
1709                         $month -= 12;
1710                         $year++;
1711                         $hebrewYear++;
1712                 }
1713
1714                 # Calculate day of year from 1 September
1715                 $dayOfYear = $day;
1716                 for ( $i = 1; $i < $month; $i++ ) {
1717                         if ( $i == 6 ) {
1718                                 # February
1719                                 $dayOfYear += 28;
1720                                 # Check if the year is leap
1721                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1722                                         $dayOfYear++;
1723                                 }
1724                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1725                                 $dayOfYear += 30;
1726                         } else {
1727                                 $dayOfYear += 31;
1728                         }
1729                 }
1730
1731                 # Calculate the start of the Hebrew year
1732                 $start = self::hebrewYearStart( $hebrewYear );
1733
1734                 # Calculate next year's start
1735                 if ( $dayOfYear <= $start ) {
1736                         # Day is before the start of the year - it is the previous year
1737                         # Next year's start
1738                         $nextStart = $start;
1739                         # Previous year
1740                         $year--;
1741                         $hebrewYear--;
1742                         # Add days since previous year's 1 September
1743                         $dayOfYear += 365;
1744                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1745                                 # Leap year
1746                                 $dayOfYear++;
1747                         }
1748                         # Start of the new (previous) year
1749                         $start = self::hebrewYearStart( $hebrewYear );
1750                 } else {
1751                         # Next year's start
1752                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1753                 }
1754
1755                 # Calculate Hebrew day of year
1756                 $hebrewDayOfYear = $dayOfYear - $start;
1757
1758                 # Difference between year's days
1759                 $diff = $nextStart - $start;
1760                 # Add 12 (or 13 for leap years) days to ignore the difference between
1761                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1762                 # difference is only about the year type
1763                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1764                         $diff += 13;
1765                 } else {
1766                         $diff += 12;
1767                 }
1768
1769                 # Check the year pattern, and is leap year
1770                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1771                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1772                 # and non-leap years
1773                 $yearPattern = $diff % 30;
1774                 # Check if leap year
1775                 $isLeap = $diff >= 30;
1776
1777                 # Calculate day in the month from number of day in the Hebrew year
1778                 # Don't check Adar - if the day is not in Adar, we will stop before;
1779                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1780                 $hebrewDay = $hebrewDayOfYear;
1781                 $hebrewMonth = 1;
1782                 $days = 0;
1783                 while ( $hebrewMonth <= 12 ) {
1784                         # Calculate days in this month
1785                         if ( $isLeap && $hebrewMonth == 6 ) {
1786                                 # Adar in a leap year
1787                                 if ( $isLeap ) {
1788                                         # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1789                                         $days = 30;
1790                                         if ( $hebrewDay <= $days ) {
1791                                                 # Day in Adar I
1792                                                 $hebrewMonth = 13;
1793                                         } else {
1794                                                 # Subtract the days of Adar I
1795                                                 $hebrewDay -= $days;
1796                                                 # Try Adar II
1797                                                 $days = 29;
1798                                                 if ( $hebrewDay <= $days ) {
1799                                                         # Day in Adar II
1800                                                         $hebrewMonth = 14;
1801                                                 }
1802                                         }
1803                                 }
1804                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1805                                 # Cheshvan in a complete year (otherwise as the rule below)
1806                                 $days = 30;
1807                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1808                                 # Kislev in an incomplete year (otherwise as the rule below)
1809                                 $days = 29;
1810                         } else {
1811                                 # Odd months have 30 days, even have 29
1812                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1813                         }
1814                         if ( $hebrewDay <= $days ) {
1815                                 # In the current month
1816                                 break;
1817                         } else {
1818                                 # Subtract the days of the current month
1819                                 $hebrewDay -= $days;
1820                                 # Try in the next month
1821                                 $hebrewMonth++;
1822                         }
1823                 }
1824
1825                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1826         }
1827
1828         /**
1829          * This calculates the Hebrew year start, as days since 1 September.
1830          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1831          * Used for Hebrew date.
1832          *
1833          * @param int $year
1834          *
1835          * @return string
1836          */
1837         private static function hebrewYearStart( $year ) {
1838                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1839                 $b = intval( ( $year - 1 ) % 4 );
1840                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1841                 if ( $m < 0 ) {
1842                         $m--;
1843                 }
1844                 $Mar = intval( $m );
1845                 if ( $m < 0 ) {
1846                         $m++;
1847                 }
1848                 $m -= $Mar;
1849
1850                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1851                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1852                         $Mar++;
1853                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1854                         $Mar += 2;
1855                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1856                         $Mar++;
1857                 }
1858
1859                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1860                 return $Mar;
1861         }
1862
1863         /**
1864          * Algorithm to convert Gregorian dates to Thai solar dates,
1865          * Minguo dates or Minguo dates.
1866          *
1867          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1868          *       https://en.wikipedia.org/wiki/Minguo_calendar
1869          *       https://en.wikipedia.org/wiki/Japanese_era_name
1870          *
1871          * @param string $ts 14-character timestamp
1872          * @param string $cName Calender name
1873          * @return array Converted year, month, day
1874          */
1875         private static function tsToYear( $ts, $cName ) {
1876                 $gy = substr( $ts, 0, 4 );
1877                 $gm = substr( $ts, 4, 2 );
1878                 $gd = substr( $ts, 6, 2 );
1879
1880                 if ( !strcmp( $cName, 'thai' ) ) {
1881                         # Thai solar dates
1882                         # Add 543 years to the Gregorian calendar
1883                         # Months and days are identical
1884                         $gy_offset = $gy + 543;
1885                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1886                         # Minguo dates
1887                         # Deduct 1911 years from the Gregorian calendar
1888                         # Months and days are identical
1889                         $gy_offset = $gy - 1911;
1890                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1891                         # Nengō dates up to Meiji period
1892                         # Deduct years from the Gregorian calendar
1893                         # depending on the nengo periods
1894                         # Months and days are identical
1895                         if ( ( $gy < 1912 )
1896                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1897                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1898                         ) {
1899                                 # Meiji period
1900                                 $gy_gannen = $gy - 1868 + 1;
1901                                 $gy_offset = $gy_gannen;
1902                                 if ( $gy_gannen == 1 ) {
1903                                         $gy_offset = '元';
1904                                 }
1905                                 $gy_offset = '明治' . $gy_offset;
1906                         } elseif (
1907                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1908                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1909                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1910                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1911                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1912                         ) {
1913                                 # Taishō period
1914                                 $gy_gannen = $gy - 1912 + 1;
1915                                 $gy_offset = $gy_gannen;
1916                                 if ( $gy_gannen == 1 ) {
1917                                         $gy_offset = '元';
1918                                 }
1919                                 $gy_offset = '大正' . $gy_offset;
1920                         } elseif (
1921                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1922                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1923                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1924                         ) {
1925                                 # Shōwa period
1926                                 $gy_gannen = $gy - 1926 + 1;
1927                                 $gy_offset = $gy_gannen;
1928                                 if ( $gy_gannen == 1 ) {
1929                                         $gy_offset = '元';
1930                                 }
1931                                 $gy_offset = '昭和' . $gy_offset;
1932                         } else {
1933                                 # Heisei period
1934                                 $gy_gannen = $gy - 1989 + 1;
1935                                 $gy_offset = $gy_gannen;
1936                                 if ( $gy_gannen == 1 ) {
1937                                         $gy_offset = '元';
1938                                 }
1939                                 $gy_offset = '平成' . $gy_offset;
1940                         }
1941                 } else {
1942                         $gy_offset = $gy;
1943                 }
1944
1945                 return [ $gy_offset, $gm, $gd ];
1946         }
1947
1948         /**
1949          * Gets directionality of the first strongly directional codepoint, for embedBidi()
1950          *
1951          * This is the rule the BIDI algorithm uses to determine the directionality of
1952          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
1953          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
1954          *
1955          * TODO: Does not handle BIDI control characters inside the text.
1956          * TODO: Does not handle unallocated characters.
1957          *
1958          * @param string $text Text to test
1959          * @return null|string Directionality ('ltr' or 'rtl') or null
1960          */
1961         private static function strongDirFromContent( $text = '' ) {
1962                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
1963                         return null;
1964                 }
1965                 if ( $matches[1] === '' ) {
1966                         return 'rtl';
1967                 }
1968                 return 'ltr';
1969         }
1970
1971         /**
1972          * Roman number formatting up to 10000
1973          *
1974          * @param int $num
1975          *
1976          * @return string
1977          */
1978         static function romanNumeral( $num ) {
1979                 static $table = [
1980                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
1981                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
1982                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
1983                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
1984                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
1985                 ];
1986
1987                 $num = intval( $num );
1988                 if ( $num > 10000 || $num <= 0 ) {
1989                         return $num;
1990                 }
1991
1992                 $s = '';
1993                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
1994                         if ( $num >= $pow10 ) {
1995                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
1996                         }
1997                         $num = $num % $pow10;
1998                 }
1999                 return $s;
2000         }
2001
2002         /**
2003          * Hebrew Gematria number formatting up to 9999
2004          *
2005          * @param int $num
2006          *
2007          * @return string
2008          */
2009         static function hebrewNumeral( $num ) {
2010                 static $table = [
2011                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2012                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2013                         [ '',
2014                                 [ 'ק' ],
2015                                 [ 'ר' ],
2016                                 [ 'ש' ],
2017                                 [ 'ת' ],
2018                                 [ 'ת', 'ק' ],
2019                                 [ 'ת', 'ר' ],
2020                                 [ 'ת', 'ש' ],
2021                                 [ 'ת', 'ת' ],
2022                                 [ 'ת', 'ת', 'ק' ],
2023                                 [ 'ת', 'ת', 'ר' ],
2024                         ],
2025                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2026                 ];
2027
2028                 $num = intval( $num );
2029                 if ( $num > 9999 || $num <= 0 ) {
2030                         return $num;
2031                 }
2032
2033                 // Round thousands have special notations
2034                 if ( $num === 1000 ) {
2035                         return "א' אלף";
2036                 } elseif ( $num % 1000 === 0 ) {
2037                         return $table[0][$num / 1000] . "' אלפים";
2038                 }
2039
2040                 $letters = [];
2041
2042                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2043                         if ( $num >= $pow10 ) {
2044                                 if ( $num === 15 || $num === 16 ) {
2045                                         $letters[] = $table[0][9];
2046                                         $letters[] = $table[0][$num - 9];
2047                                         $num = 0;
2048                                 } else {
2049                                         $letters = array_merge(
2050                                                 $letters,
2051                                                 (array)$table[$i][intval( $num / $pow10 )]
2052                                         );
2053
2054                                         if ( $pow10 === 1000 ) {
2055                                                 $letters[] = "'";
2056                                         }
2057                                 }
2058                         }
2059
2060                         $num = $num % $pow10;
2061                 }
2062
2063                 $preTransformLength = count( $letters );
2064                 if ( $preTransformLength === 1 ) {
2065                         // Add geresh (single quote) to one-letter numbers
2066                         $letters[] = "'";
2067                 } else {
2068                         $lastIndex = $preTransformLength - 1;
2069                         $letters[$lastIndex] = str_replace(
2070                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2071                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2072                                 $letters[$lastIndex]
2073                         );
2074
2075                         // Add gershayim (double quote) to multiple-letter numbers,
2076                         // but exclude numbers with only one letter after the thousands
2077                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2078                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2079                                 $letters[] = "'";
2080                         } else {
2081                                 array_splice( $letters, -1, 0, '"' );
2082                         }
2083                 }
2084
2085                 return implode( $letters );
2086         }
2087
2088         /**
2089          * Used by date() and time() to adjust the time output.
2090          *
2091          * @param string $ts The time in date('YmdHis') format
2092          * @param mixed $tz Adjust the time by this amount (default false, mean we
2093          *   get user timecorrection setting)
2094          * @return int
2095          */
2096         public function userAdjust( $ts, $tz = false ) {
2097                 global $wgUser, $wgLocalTZoffset;
2098
2099                 if ( $tz === false ) {
2100                         $tz = $wgUser->getOption( 'timecorrection' );
2101                 }
2102
2103                 $data = explode( '|', $tz, 3 );
2104
2105                 if ( $data[0] == 'ZoneInfo' ) {
2106                         try {
2107                                 $userTZ = new DateTimeZone( $data[2] );
2108                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2109                                 $date->setTimezone( $userTZ );
2110                                 return $date->format( 'YmdHis' );
2111                         } catch ( Exception $e ) {
2112                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2113                                 $data[0] = 'Offset';
2114                         }
2115                 }
2116
2117                 if ( $data[0] == 'System' || $tz == '' ) {
2118                         # Global offset in minutes.
2119                         $minDiff = $wgLocalTZoffset;
2120                 } elseif ( $data[0] == 'Offset' ) {
2121                         $minDiff = intval( $data[1] );
2122                 } else {
2123                         $data = explode( ':', $tz );
2124                         if ( count( $data ) == 2 ) {
2125                                 $data[0] = intval( $data[0] );
2126                                 $data[1] = intval( $data[1] );
2127                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2128                                 if ( $data[0] < 0 ) {
2129                                         $minDiff = -$minDiff;
2130                                 }
2131                         } else {
2132                                 $minDiff = intval( $data[0] ) * 60;
2133                         }
2134                 }
2135
2136                 # No difference ? Return time unchanged
2137                 if ( 0 == $minDiff ) {
2138                         return $ts;
2139                 }
2140
2141                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2142                 # Generate an adjusted date; take advantage of the fact that mktime
2143                 # will normalize out-of-range values so we don't have to split $minDiff
2144                 # into hours and minutes.
2145                 $t = mktime( (
2146                         (int)substr( $ts, 8, 2 ) ), # Hours
2147                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2148                         (int)substr( $ts, 12, 2 ), # Seconds
2149                         (int)substr( $ts, 4, 2 ), # Month
2150                         (int)substr( $ts, 6, 2 ), # Day
2151                         (int)substr( $ts, 0, 4 ) ); # Year
2152
2153                 $date = date( 'YmdHis', $t );
2154                 Wikimedia\restoreWarnings();
2155
2156                 return $date;
2157         }
2158
2159         /**
2160          * This is meant to be used by time(), date(), and timeanddate() to get
2161          * the date preference they're supposed to use, it should be used in
2162          * all children.
2163          *
2164          *     function timeanddate([...], $format = true) {
2165          *       $datePreference = $this->dateFormat($format);
2166          *       [...]
2167          *     }
2168          *
2169          * @param int|string|bool $usePrefs If true, the user's preference is used
2170          *   if false, the site/language default is used
2171          *   if int/string, assumed to be a format.
2172          * @return string
2173          */
2174         function dateFormat( $usePrefs = true ) {
2175                 global $wgUser;
2176
2177                 if ( is_bool( $usePrefs ) ) {
2178                         if ( $usePrefs ) {
2179                                 $datePreference = $wgUser->getDatePreference();
2180                         } else {
2181                                 $datePreference = (string)User::getDefaultOption( 'date' );
2182                         }
2183                 } else {
2184                         $datePreference = (string)$usePrefs;
2185                 }
2186
2187                 // return int
2188                 if ( $datePreference == '' ) {
2189                         return 'default';
2190                 }
2191
2192                 return $datePreference;
2193         }
2194
2195         /**
2196          * Get a format string for a given type and preference
2197          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2198          * @param string $pref The format name as it appears in Messages*.php under
2199          *  $datePreferences.
2200          *
2201          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2202          *
2203          * @return string
2204          */
2205         function getDateFormatString( $type, $pref ) {
2206                 $wasDefault = false;
2207                 if ( $pref == 'default' ) {
2208                         $wasDefault = true;
2209                         $pref = $this->getDefaultDateFormat();
2210                 }
2211
2212                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2213                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2214
2215                         if ( $type === 'pretty' && $df === null ) {
2216                                 $df = $this->getDateFormatString( 'date', $pref );
2217                         }
2218
2219                         if ( !$wasDefault && $df === null ) {
2220                                 $pref = $this->getDefaultDateFormat();
2221                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2222                         }
2223
2224                         $this->dateFormatStrings[$type][$pref] = $df;
2225                 }
2226                 return $this->dateFormatStrings[$type][$pref];
2227         }
2228
2229         /**
2230          * @param string $ts The time format which needs to be turned into a
2231          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2232          * @param bool $adj Whether to adjust the time output according to the
2233          *   user configured offset ($timecorrection)
2234          * @param mixed $format True to use user's date format preference
2235          * @param string|bool $timecorrection The time offset as returned by
2236          *   validateTimeZone() in Special:Preferences
2237          * @return string
2238          */
2239         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2240                 $ts = wfTimestamp( TS_MW, $ts );
2241                 if ( $adj ) {
2242                         $ts = $this->userAdjust( $ts, $timecorrection );
2243                 }
2244                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2245                 return $this->sprintfDate( $df, $ts );
2246         }
2247
2248         /**
2249          * @param string $ts The time format which needs to be turned into a
2250          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2251          * @param bool $adj Whether to adjust the time output according to the
2252          *   user configured offset ($timecorrection)
2253          * @param mixed $format True to use user's date format preference
2254          * @param string|bool $timecorrection The time offset as returned by
2255          *   validateTimeZone() in Special:Preferences
2256          * @return string
2257          */
2258         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2259                 $ts = wfTimestamp( TS_MW, $ts );
2260                 if ( $adj ) {
2261                         $ts = $this->userAdjust( $ts, $timecorrection );
2262                 }
2263                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2264                 return $this->sprintfDate( $df, $ts );
2265         }
2266
2267         /**
2268          * @param string $ts The time format which needs to be turned into a
2269          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2270          * @param bool $adj Whether to adjust the time output according to the
2271          *   user configured offset ($timecorrection)
2272          * @param mixed $format What format to return, if it's false output the
2273          *   default one (default true)
2274          * @param string|bool $timecorrection The time offset as returned by
2275          *   validateTimeZone() in Special:Preferences
2276          * @return string
2277          */
2278         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2279                 $ts = wfTimestamp( TS_MW, $ts );
2280                 if ( $adj ) {
2281                         $ts = $this->userAdjust( $ts, $timecorrection );
2282                 }
2283                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2284                 return $this->sprintfDate( $df, $ts );
2285         }
2286
2287         /**
2288          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2289          *
2290          * @since 1.20
2291          *
2292          * @param int $seconds The amount of seconds.
2293          * @param array $chosenIntervals The intervals to enable.
2294          *
2295          * @return string
2296          */
2297         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2298                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2299
2300                 $segments = [];
2301
2302                 foreach ( $intervals as $intervalName => $intervalValue ) {
2303                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2304                         // duration-years, duration-decades, duration-centuries, duration-millennia
2305                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2306                         $segments[] = $message->inLanguage( $this )->escaped();
2307                 }
2308
2309                 return $this->listToText( $segments );
2310         }
2311
2312         /**
2313          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2314          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2315          *
2316          * @since 1.20
2317          *
2318          * @param int $seconds The amount of seconds.
2319          * @param array $chosenIntervals The intervals to enable.
2320          *
2321          * @return array
2322          */
2323         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2324                 if ( empty( $chosenIntervals ) ) {
2325                         $chosenIntervals = [
2326                                 'millennia',
2327                                 'centuries',
2328                                 'decades',
2329                                 'years',
2330                                 'days',
2331                                 'hours',
2332                                 'minutes',
2333                                 'seconds'
2334                         ];
2335                 }
2336
2337                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2338                 $sortedNames = array_keys( $intervals );
2339                 $smallestInterval = array_pop( $sortedNames );
2340
2341                 $segments = [];
2342
2343                 foreach ( $intervals as $name => $length ) {
2344                         $value = floor( $seconds / $length );
2345
2346                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2347                                 $seconds -= $value * $length;
2348                                 $segments[$name] = $value;
2349                         }
2350                 }
2351
2352                 return $segments;
2353         }
2354
2355         /**
2356          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2357          *
2358          * @param string $type Can be 'date', 'time' or 'both'
2359          * @param string $ts The time format which needs to be turned into a
2360          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2361          * @param User $user User object used to get preferences for timezone and format
2362          * @param array $options Array, can contain the following keys:
2363          *   - 'timecorrection': time correction, can have the following values:
2364          *     - true: use user's preference
2365          *     - false: don't use time correction
2366          *     - int: value of time correction in minutes
2367          *   - 'format': format to use, can have the following values:
2368          *     - true: use user's preference
2369          *     - false: use default preference
2370          *     - string: format to use
2371          * @since 1.19
2372          * @return string
2373          */
2374         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2375                 $ts = wfTimestamp( TS_MW, $ts );
2376                 $options += [ 'timecorrection' => true, 'format' => true ];
2377                 if ( $options['timecorrection'] !== false ) {
2378                         if ( $options['timecorrection'] === true ) {
2379                                 $offset = $user->getOption( 'timecorrection' );
2380                         } else {
2381                                 $offset = $options['timecorrection'];
2382                         }
2383                         $ts = $this->userAdjust( $ts, $offset );
2384                 }
2385                 if ( $options['format'] === true ) {
2386                         $format = $user->getDatePreference();
2387                 } else {
2388                         $format = $options['format'];
2389                 }
2390                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2391                 return $this->sprintfDate( $df, $ts );
2392         }
2393
2394         /**
2395          * Get the formatted date for the given timestamp and formatted for
2396          * the given user.
2397          *
2398          * @param mixed $ts Mixed: the time format which needs to be turned into a
2399          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2400          * @param User $user User object used to get preferences for timezone and format
2401          * @param array $options Array, can contain the following keys:
2402          *   - 'timecorrection': time correction, can have the following values:
2403          *     - true: use user's preference
2404          *     - false: don't use time correction
2405          *     - int: value of time correction in minutes
2406          *   - 'format': format to use, can have the following values:
2407          *     - true: use user's preference
2408          *     - false: use default preference
2409          *     - string: format to use
2410          * @since 1.19
2411          * @return string
2412          */
2413         public function userDate( $ts, User $user, array $options = [] ) {
2414                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2415         }
2416
2417         /**
2418          * Get the formatted time for the given timestamp and formatted for
2419          * the given user.
2420          *
2421          * @param mixed $ts The time format which needs to be turned into a
2422          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2423          * @param User $user User object used to get preferences for timezone and format
2424          * @param array $options Array, can contain the following keys:
2425          *   - 'timecorrection': time correction, can have the following values:
2426          *     - true: use user's preference
2427          *     - false: don't use time correction
2428          *     - int: value of time correction in minutes
2429          *   - 'format': format to use, can have the following values:
2430          *     - true: use user's preference
2431          *     - false: use default preference
2432          *     - string: format to use
2433          * @since 1.19
2434          * @return string
2435          */
2436         public function userTime( $ts, User $user, array $options = [] ) {
2437                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2438         }
2439
2440         /**
2441          * Get the formatted date and time for the given timestamp and formatted for
2442          * the given user.
2443          *
2444          * @param mixed $ts The time format which needs to be turned into a
2445          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2446          * @param User $user User object used to get preferences for timezone and format
2447          * @param array $options Array, can contain the following keys:
2448          *   - 'timecorrection': time correction, can have the following values:
2449          *     - true: use user's preference
2450          *     - false: don't use time correction
2451          *     - int: value of time correction in minutes
2452          *   - 'format': format to use, can have the following values:
2453          *     - true: use user's preference
2454          *     - false: use default preference
2455          *     - string: format to use
2456          * @since 1.19
2457          * @return string
2458          */
2459         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2460                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2461         }
2462
2463         /**
2464          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2465          *
2466          * Determine the difference between the timestamp and the current time, and
2467          * generate a readable timestamp by returning "<N> <units> ago", where the
2468          * largest possible unit is used.
2469          *
2470          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2471          *
2472          * @param MWTimestamp $time
2473          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2474          * @param User|null $user User the timestamp is being generated for
2475          *  (or null to use main context's user)
2476          * @return string Formatted timestamp
2477          */
2478         public function getHumanTimestamp(
2479                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2480         ) {
2481                 if ( $relativeTo === null ) {
2482                         $relativeTo = new MWTimestamp();
2483                 }
2484                 if ( $user === null ) {
2485                         $user = RequestContext::getMain()->getUser();
2486                 }
2487
2488                 // Adjust for the user's timezone.
2489                 $offsetThis = $time->offsetForUser( $user );
2490                 $offsetRel = $relativeTo->offsetForUser( $user );
2491
2492                 $ts = '';
2493                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2494                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2495                 }
2496
2497                 // Reset the timezone on the objects.
2498                 $time->timestamp->sub( $offsetThis );
2499                 $relativeTo->timestamp->sub( $offsetRel );
2500
2501                 return $ts;
2502         }
2503
2504         /**
2505          * Convert an MWTimestamp into a pretty human-readable timestamp using
2506          * the given user preferences and relative base time.
2507          *
2508          * @see Language::getHumanTimestamp
2509          * @param MWTimestamp $ts Timestamp to prettify
2510          * @param MWTimestamp $relativeTo Base timestamp
2511          * @param User $user User preferences to use
2512          * @return string Human timestamp
2513          * @since 1.26
2514          */
2515         private function getHumanTimestampInternal(
2516                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2517         ) {
2518                 $diff = $ts->diff( $relativeTo );
2519                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2520                         (int)$relativeTo->timestamp->format( 'w' ) );
2521                 $days = $diff->days ?: (int)$diffDay;
2522                 if ( $diff->invert || $days > 5
2523                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2524                 ) {
2525                         // Timestamps are in different years: use full timestamp
2526                         // Also do full timestamp for future dates
2527                         /**
2528                          * @todo FIXME: Add better handling of future timestamps.
2529                          */
2530                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2531                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2532                 } elseif ( $days > 5 ) {
2533                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2534                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2535                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2536                 } elseif ( $days > 1 ) {
2537                         // Timestamp within the past week: show the day of the week and time
2538                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2539                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2540                         // Messages:
2541                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2542                         $ts = wfMessage( "$weekday-at" )
2543                                 ->inLanguage( $this )
2544                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2545                                 ->text();
2546                 } elseif ( $days == 1 ) {
2547                         // Timestamp was yesterday: say 'yesterday' and the time.
2548                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2549                         $ts = wfMessage( 'yesterday-at' )
2550                                 ->inLanguage( $this )
2551                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2552                                 ->text();
2553                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2554                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2555                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2556                         $ts = wfMessage( 'today-at' )
2557                                 ->inLanguage( $this )
2558                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2559                                 ->text();
2560
2561                 // From here on in, the timestamp was soon enough ago so that we can simply say
2562                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2563                 } elseif ( $diff->h == 1 ) {
2564                         // Less than 90 minutes, but more than an hour ago.
2565                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2566                 } elseif ( $diff->i >= 1 ) {
2567                         // A few minutes ago.
2568                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2569                 } elseif ( $diff->s >= 30 ) {
2570                         // Less than a minute, but more than 30 sec ago.
2571                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2572                 } else {
2573                         // Less than 30 seconds ago.
2574                         $ts = wfMessage( 'just-now' )->text();
2575                 }
2576
2577                 return $ts;
2578         }
2579
2580         /**
2581          * @param string $key
2582          * @return string|null
2583          */
2584         public function getMessage( $key ) {
2585                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2586         }
2587
2588         /**
2589          * @return array
2590          */
2591         function getAllMessages() {
2592                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2593         }
2594
2595         /**
2596          * @param string $in
2597          * @param string $out
2598          * @param string $string
2599          * @return string
2600          */
2601         public function iconv( $in, $out, $string ) {
2602                 # Even with //IGNORE iconv can whine about illegal characters in
2603                 # *input* string. We just ignore those too.
2604                 # REF: https://bugs.php.net/bug.php?id=37166
2605                 # REF: https://phabricator.wikimedia.org/T18885
2606                 Wikimedia\suppressWarnings();
2607                 $text = iconv( $in, $out . '//IGNORE', $string );
2608                 Wikimedia\restoreWarnings();
2609                 return $text;
2610         }
2611
2612         // callback functions for ucwords(), ucwordbreaks()
2613
2614         /**
2615          * @param array $matches
2616          * @return mixed|string
2617          */
2618         function ucwordbreaksCallbackAscii( $matches ) {
2619                 return $this->ucfirst( $matches[1] );
2620         }
2621
2622         /**
2623          * @param array $matches
2624          * @return string
2625          */
2626         function ucwordbreaksCallbackMB( $matches ) {
2627                 return mb_strtoupper( $matches[0] );
2628         }
2629
2630         /**
2631          * @param array $matches
2632          * @return string
2633          */
2634         function ucwordsCallbackMB( $matches ) {
2635                 return mb_strtoupper( $matches[0] );
2636         }
2637
2638         /**
2639          * Make a string's first character uppercase
2640          *
2641          * @param string $str
2642          *
2643          * @return string
2644          */
2645         public function ucfirst( $str ) {
2646                 $o = ord( $str );
2647                 if ( $o < 96 ) { // if already uppercase...
2648                         return $str;
2649                 } elseif ( $o < 128 ) {
2650                         return ucfirst( $str ); // use PHP's ucfirst()
2651                 } else {
2652                         // fall back to more complex logic in case of multibyte strings
2653                         return $this->uc( $str, true );
2654                 }
2655         }
2656
2657         /**
2658          * Convert a string to uppercase
2659          *
2660          * @param string $str
2661          * @param bool $first
2662          *
2663          * @return string
2664          */
2665         public function uc( $str, $first = false ) {
2666                 if ( $first ) {
2667                         if ( $this->isMultibyte( $str ) ) {
2668                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2669                         } else {
2670                                 return ucfirst( $str );
2671                         }
2672                 } else {
2673                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2674                 }
2675         }
2676
2677         /**
2678          * @param string $str
2679          * @return mixed|string
2680          */
2681         function lcfirst( $str ) {
2682                 $o = ord( $str );
2683                 if ( !$o ) {
2684                         return strval( $str );
2685                 } elseif ( $o >= 128 ) {
2686                         return $this->lc( $str, true );
2687                 } elseif ( $o > 96 ) {
2688                         return $str;
2689                 } else {
2690                         $str[0] = strtolower( $str[0] );
2691                         return $str;
2692                 }
2693         }
2694
2695         /**
2696          * @param string $str
2697          * @param bool $first
2698          * @return mixed|string
2699          */
2700         function lc( $str, $first = false ) {
2701                 if ( $first ) {
2702                         if ( $this->isMultibyte( $str ) ) {
2703                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2704                         } else {
2705                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2706                         }
2707                 } else {
2708                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2709                 }
2710         }
2711
2712         /**
2713          * @param string $str
2714          * @return bool
2715          */
2716         function isMultibyte( $str ) {
2717                 return strlen( $str ) !== mb_strlen( $str );
2718         }
2719
2720         /**
2721          * @param string $str
2722          * @return mixed|string
2723          */
2724         function ucwords( $str ) {
2725                 if ( $this->isMultibyte( $str ) ) {
2726                         $str = $this->lc( $str );
2727
2728                         // regexp to find first letter in each word (i.e. after each space)
2729                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2730
2731                         // function to use to capitalize a single char
2732                         return preg_replace_callback(
2733                                 $replaceRegexp,
2734                                 [ $this, 'ucwordsCallbackMB' ],
2735                                 $str
2736                         );
2737                 } else {
2738                         return ucwords( strtolower( $str ) );
2739                 }
2740         }
2741
2742         /**
2743          * capitalize words at word breaks
2744          *
2745          * @param string $str
2746          * @return mixed
2747          */
2748         function ucwordbreaks( $str ) {
2749                 if ( $this->isMultibyte( $str ) ) {
2750                         $str = $this->lc( $str );
2751
2752                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2753                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2754
2755                         // find first letter after word break
2756                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2757                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2758
2759                         return preg_replace_callback(
2760                                 $replaceRegexp,
2761                                 [ $this, 'ucwordbreaksCallbackMB' ],
2762                                 $str
2763                         );
2764                 } else {
2765                         return preg_replace_callback(
2766                                 '/\b([\w\x80-\xff]+)\b/',
2767                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2768                                 $str
2769                         );
2770                 }
2771         }
2772
2773         /**
2774          * Return a case-folded representation of $s
2775          *
2776          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2777          * and $s2 are the same except for the case of their characters. It is not
2778          * necessary for the value returned to make sense when displayed.
2779          *
2780          * Do *not* perform any other normalisation in this function. If a caller
2781          * uses this function when it should be using a more general normalisation
2782          * function, then fix the caller.
2783          *
2784          * @param string $s
2785          *
2786          * @return string
2787          */
2788         function caseFold( $s ) {
2789                 return $this->uc( $s );
2790         }
2791
2792         /**
2793          * @param string $s
2794          * @return string
2795          * @throws MWException
2796          */
2797         function checkTitleEncoding( $s ) {
2798                 if ( is_array( $s ) ) {
2799                         throw new MWException( 'Given array to checkTitleEncoding.' );
2800                 }
2801                 if ( StringUtils::isUtf8( $s ) ) {
2802                         return $s;
2803                 }
2804
2805                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2806         }
2807
2808         /**
2809          * @return array
2810          */
2811         function fallback8bitEncoding() {
2812                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2813         }
2814
2815         /**
2816          * Most writing systems use whitespace to break up words.
2817          * Some languages such as Chinese don't conventionally do this,
2818          * which requires special handling when breaking up words for
2819          * searching etc.
2820          *
2821          * @return bool
2822          */
2823         function hasWordBreaks() {
2824                 return true;
2825         }
2826
2827         /**
2828          * Some languages such as Chinese require word segmentation,
2829          * Specify such segmentation when overridden in derived class.
2830          *
2831          * @param string $string
2832          * @return string
2833          */
2834         function segmentByWord( $string ) {
2835                 return $string;
2836         }
2837
2838         /**
2839          * Some languages have special punctuation need to be normalized.
2840          * Make such changes here.
2841          *
2842          * @param string $string
2843          * @return string
2844          */
2845         function normalizeForSearch( $string ) {
2846                 return self::convertDoubleWidth( $string );
2847         }
2848
2849         /**
2850          * convert double-width roman characters to single-width.
2851          * range: ff00-ff5f ~= 0020-007f
2852          *
2853          * @param string $string
2854          *
2855          * @return string
2856          */
2857         protected static function convertDoubleWidth( $string ) {
2858                 static $full = null;
2859                 static $half = null;
2860
2861                 if ( $full === null ) {
2862                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2863                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2864                         $full = str_split( $fullWidth, 3 );
2865                         $half = str_split( $halfWidth );
2866                 }
2867
2868                 $string = str_replace( $full, $half, $string );
2869                 return $string;
2870         }
2871
2872         /**
2873          * @param string $string
2874          * @param string $pattern
2875          * @return string
2876          */
2877         protected static function insertSpace( $string, $pattern ) {
2878                 $string = preg_replace( $pattern, " $1 ", $string );
2879                 $string = preg_replace( '/ +/', ' ', $string );
2880                 return $string;
2881         }
2882
2883         /**
2884          * @param array $termsArray
2885          * @return array
2886          */
2887         function convertForSearchResult( $termsArray ) {
2888                 # some languages, e.g. Chinese, need to do a conversion
2889                 # in order for search results to be displayed correctly
2890                 return $termsArray;
2891         }
2892
2893         /**
2894          * Get the first character of a string.
2895          *
2896          * @param string $s
2897          * @return string
2898          */
2899         function firstChar( $s ) {
2900                 $matches = [];
2901                 preg_match(
2902                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2903                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2904                         $s,
2905                         $matches
2906                 );
2907
2908                 if ( isset( $matches[1] ) ) {
2909                         if ( strlen( $matches[1] ) != 3 ) {
2910                                 return $matches[1];
2911                         }
2912
2913                         // Break down Hangul syllables to grab the first jamo
2914                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2915                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2916                                 return $matches[1];
2917                         } elseif ( $code < 0xb098 ) {
2918                                 return "\u{3131}";
2919                         } elseif ( $code < 0xb2e4 ) {
2920                                 return "\u{3134}";
2921                         } elseif ( $code < 0xb77c ) {
2922                                 return "\u{3137}";
2923                         } elseif ( $code < 0xb9c8 ) {
2924                                 return "\u{3139}";
2925                         } elseif ( $code < 0xbc14 ) {
2926                                 return "\u{3141}";
2927                         } elseif ( $code < 0xc0ac ) {
2928                                 return "\u{3142}";
2929                         } elseif ( $code < 0xc544 ) {
2930                                 return "\u{3145}";
2931                         } elseif ( $code < 0xc790 ) {
2932                                 return "\u{3147}";
2933                         } elseif ( $code < 0xcc28 ) {
2934                                 return "\u{3148}";
2935                         } elseif ( $code < 0xce74 ) {
2936                                 return "\u{314A}";
2937                         } elseif ( $code < 0xd0c0 ) {
2938                                 return "\u{314B}";
2939                         } elseif ( $code < 0xd30c ) {
2940                                 return "\u{314C}";
2941                         } elseif ( $code < 0xd558 ) {
2942                                 return "\u{314D}";
2943                         } else {
2944                                 return "\u{314E}";
2945                         }
2946                 } else {
2947                         return '';
2948                 }
2949         }
2950
2951         /**
2952          * @deprecated No-op since 1.28
2953          */
2954         function initEncoding() {
2955                 wfDeprecated( __METHOD__, '1.28' );
2956                 // No-op.
2957         }
2958
2959         /**
2960          * @param string $s
2961          * @return string
2962          * @deprecated No-op since 1.28
2963          */
2964         function recodeForEdit( $s ) {
2965                 wfDeprecated( __METHOD__, '1.28' );
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                 wfDeprecated( __METHOD__, '1.28' );
2976                 return $s;
2977         }
2978
2979         /**
2980          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
2981          * also cleans up certain backwards-compatible sequences, converting them
2982          * to the modern Unicode equivalent.
2983          *
2984          * This is language-specific for performance reasons only.
2985          *
2986          * @param string $s
2987          *
2988          * @return string
2989          */
2990         function normalize( $s ) {
2991                 global $wgAllUnicodeFixes;
2992                 $s = UtfNormal\Validator::cleanUp( $s );
2993                 if ( $wgAllUnicodeFixes ) {
2994                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
2995                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
2996                 }
2997
2998                 return $s;
2999         }
3000
3001         /**
3002          * Transform a string using serialized data stored in the given file (which
3003          * must be in the serialized subdirectory of $IP). The file contains pairs
3004          * mapping source characters to destination characters.
3005          *
3006          * The data is cached in process memory. This will go faster if you have the
3007          * FastStringSearch extension.
3008          *
3009          * @param string $file
3010          * @param string $string
3011          *
3012          * @throws MWException
3013          * @return string
3014          */
3015         protected function transformUsingPairFile( $file, $string ) {
3016                 if ( !isset( $this->transformData[$file] ) ) {
3017                         global $IP;
3018                         $data = require "$IP/languages/data/{$file}";
3019                         $this->transformData[$file] = new ReplacementArray( $data );
3020                 }
3021                 return $this->transformData[$file]->replace( $string );
3022         }
3023
3024         /**
3025          * For right-to-left language support
3026          *
3027          * @return bool
3028          */
3029         function isRTL() {
3030                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3031         }
3032
3033         /**
3034          * Return the correct HTML 'dir' attribute value for this language.
3035          * @return string
3036          */
3037         function getDir() {
3038                 return $this->isRTL() ? 'rtl' : 'ltr';
3039         }
3040
3041         /**
3042          * Return 'left' or 'right' as appropriate alignment for line-start
3043          * for this language's text direction.
3044          *
3045          * Should be equivalent to CSS3 'start' text-align value....
3046          *
3047          * @return string
3048          */
3049         function alignStart() {
3050                 return $this->isRTL() ? 'right' : 'left';
3051         }
3052
3053         /**
3054          * Return 'right' or 'left' as appropriate alignment for line-end
3055          * for this language's text direction.
3056          *
3057          * Should be equivalent to CSS3 'end' text-align value....
3058          *
3059          * @return string
3060          */
3061         function alignEnd() {
3062                 return $this->isRTL() ? 'left' : 'right';
3063         }
3064
3065         /**
3066          * A hidden direction mark (LRM or RLM), depending on the language direction.
3067          * Unlike getDirMark(), this function returns the character as an HTML entity.
3068          * This function should be used when the output is guaranteed to be HTML,
3069          * because it makes the output HTML source code more readable. When
3070          * the output is plain text or can be escaped, getDirMark() should be used.
3071          *
3072          * @param bool $opposite Get the direction mark opposite to your language
3073          * @return string
3074          * @since 1.20
3075          */
3076         function getDirMarkEntity( $opposite = false ) {
3077                 if ( $opposite ) {
3078                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3079                 }
3080                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3081         }
3082
3083         /**
3084          * A hidden direction mark (LRM or RLM), depending on the language direction.
3085          * This function produces them as invisible Unicode characters and
3086          * the output may be hard to read and debug, so it should only be used
3087          * when the output is plain text or can be escaped. When the output is
3088          * HTML, use getDirMarkEntity() instead.
3089          *
3090          * @param bool $opposite Get the direction mark opposite to your language
3091          * @return string
3092          */
3093         function getDirMark( $opposite = false ) {
3094                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3095                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3096                 if ( $opposite ) {
3097                         return $this->isRTL() ? $lrm : $rlm;
3098                 }
3099                 return $this->isRTL() ? $rlm : $lrm;
3100         }
3101
3102         /**
3103          * @return array
3104          */
3105         function capitalizeAllNouns() {
3106                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3107         }
3108
3109         /**
3110          * An arrow, depending on the language direction.
3111          *
3112          * @param string $direction The direction of the arrow: forwards (default),
3113          *   backwards, left, right, up, down.
3114          * @return string
3115          */
3116         function getArrow( $direction = 'forwards' ) {
3117                 switch ( $direction ) {
3118                         case 'forwards':
3119                                 return $this->isRTL() ? '←' : '→';
3120                         case 'backwards':
3121                                 return $this->isRTL() ? '→' : '←';
3122                         case 'left':
3123                                 return '←';
3124                         case 'right':
3125                                 return '→';
3126                         case 'up':
3127                                 return '↑';
3128                         case 'down':
3129                                 return '↓';
3130                 }
3131         }
3132
3133         /**
3134          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3135          *
3136          * @return bool
3137          */
3138         function linkPrefixExtension() {
3139                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3140         }
3141
3142         /**
3143          * Get all magic words from cache.
3144          * @return array
3145          */
3146         function getMagicWords() {
3147                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3148         }
3149
3150         /**
3151          * Run the LanguageGetMagic hook once.
3152          */
3153         protected function doMagicHook() {
3154                 if ( $this->mMagicHookDone ) {
3155                         return;
3156                 }
3157                 $this->mMagicHookDone = true;
3158                 Hooks::run( 'LanguageGetMagic', [ &$this->mMagicExtensions, $this->getCode() ], '1.16' );
3159         }
3160
3161         /**
3162          * Fill a MagicWord object with data from here
3163          *
3164          * @param MagicWord $mw
3165          */
3166         function getMagic( $mw ) {
3167                 // Saves a function call
3168                 if ( !$this->mMagicHookDone ) {
3169                         $this->doMagicHook();
3170                 }
3171
3172                 if ( isset( $this->mMagicExtensions[$mw->mId] ) ) {
3173                         $rawEntry = $this->mMagicExtensions[$mw->mId];
3174                 } else {
3175                         $rawEntry = self::$dataCache->getSubitem(
3176                                 $this->mCode, 'magicWords', $mw->mId );
3177                 }
3178
3179                 if ( !is_array( $rawEntry ) ) {
3180                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3181                 } else {
3182                         $mw->mCaseSensitive = $rawEntry[0];
3183                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3184                 }
3185         }
3186
3187         /**
3188          * Add magic words to the extension array
3189          *
3190          * @param array $newWords
3191          */
3192         function addMagicWordsByLang( $newWords ) {
3193                 $fallbackChain = $this->getFallbackLanguages();
3194                 $fallbackChain = array_reverse( $fallbackChain );
3195                 foreach ( $fallbackChain as $code ) {
3196                         if ( isset( $newWords[$code] ) ) {
3197                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3198                         }
3199                 }
3200         }
3201
3202         /**
3203          * Get special page names, as an associative array
3204          *   canonical name => array of valid names, including aliases
3205          * @return array
3206          */
3207         function getSpecialPageAliases() {
3208                 // Cache aliases because it may be slow to load them
3209                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3210                         // Initialise array
3211                         $this->mExtendedSpecialPageAliases =
3212                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3213                         Hooks::run( 'LanguageGetSpecialPageAliases',
3214                                 [ &$this->mExtendedSpecialPageAliases, $this->getCode() ], '1.16' );
3215                 }
3216
3217                 return $this->mExtendedSpecialPageAliases;
3218         }
3219
3220         /**
3221          * Italic is unsuitable for some languages
3222          *
3223          * @param string $text The text to be emphasized.
3224          * @return string
3225          */
3226         function emphasize( $text ) {
3227                 return "<em>$text</em>";
3228         }
3229
3230         /**
3231          * Normally we output all numbers in plain en_US style, that is
3232          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3233          * point twohundredthirtyfive. However this is not suitable for all
3234          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3235          * Icelandic just want to use commas instead of dots, and dots instead
3236          * of commas like "293.291,235".
3237          *
3238          * An example of this function being called:
3239          * <code>
3240          * wfMessage( 'message' )->numParams( $num )->text()
3241          * </code>
3242          *
3243          * See $separatorTransformTable on MessageIs.php for
3244          * the , => . and . => , implementation.
3245          *
3246          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3247          * @param int|float $number The string to be formatted, should be an integer
3248          *   or a floating point number.
3249          * @param bool $nocommafy Set to true for special numbers like dates
3250          * @return string
3251          */
3252         public function formatNum( $number, $nocommafy = false ) {
3253                 global $wgTranslateNumerals;
3254                 if ( !$nocommafy ) {
3255                         $number = $this->commafy( $number );
3256                         $s = $this->separatorTransformTable();
3257                         if ( $s ) {
3258                                 $number = strtr( $number, $s );
3259                         }
3260                 }
3261
3262                 if ( $wgTranslateNumerals ) {
3263                         $s = $this->digitTransformTable();
3264                         if ( $s ) {
3265                                 $number = strtr( $number, $s );
3266                         }
3267                 }
3268
3269                 return (string)$number;
3270         }
3271
3272         /**
3273          * Front-end for non-commafied formatNum
3274          *
3275          * @param int|float $number The string to be formatted, should be an integer
3276          *        or a floating point number.
3277          * @since 1.21
3278          * @return string
3279          */
3280         public function formatNumNoSeparators( $number ) {
3281                 return $this->formatNum( $number, true );
3282         }
3283
3284         /**
3285          * @param string $number
3286          * @return string
3287          */
3288         public function parseFormattedNumber( $number ) {
3289                 $s = $this->digitTransformTable();
3290                 if ( $s ) {
3291                         // eliminate empty array values such as ''. (T66347)
3292                         $s = array_filter( $s );
3293                         $number = strtr( $number, array_flip( $s ) );
3294                 }
3295
3296                 $s = $this->separatorTransformTable();
3297                 if ( $s ) {
3298                         // eliminate empty array values such as ''. (T66347)
3299                         $s = array_filter( $s );
3300                         $number = strtr( $number, array_flip( $s ) );
3301                 }
3302
3303                 $number = strtr( $number, [ ',' => '' ] );
3304                 return $number;
3305         }
3306
3307         /**
3308          * Adds commas to a given number
3309          * @since 1.19
3310          * @param mixed $number
3311          * @return string
3312          */
3313         function commafy( $number ) {
3314                 $digitGroupingPattern = $this->digitGroupingPattern();
3315                 $minimumGroupingDigits = $this->minimumGroupingDigits();
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                         // In some languages it's conventional not to insert a thousands separator
3323                         // in numbers that are four digits long (1000-9999).
3324                         if ( $minimumGroupingDigits ) {
3325                                 // Number of '#' characters after last comma in the grouping pattern.
3326                                 // The pattern is hardcoded here, but this would vary for different patterns.
3327                                 $primaryGroupingSize = 3;
3328                                 // Maximum length of a number to suppress digit grouping for.
3329                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3330                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3331                                         return $number;
3332                                 }
3333                         }
3334                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3335                 } else {
3336                         // Ref: http://cldr.unicode.org/translation/number-patterns
3337                         $sign = "";
3338                         if ( intval( $number ) < 0 ) {
3339                                 // For negative numbers apply the algorithm like positive number and add sign.
3340                                 $sign = "-";
3341                                 $number = substr( $number, 1 );
3342                         }
3343                         $integerPart = [];
3344                         $decimalPart = [];
3345                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3346                         preg_match( "/\d+/", $number, $integerPart );
3347                         preg_match( "/\.\d*/", $number, $decimalPart );
3348                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3349                         if ( $groupedNumber === $number ) {
3350                                 // the string does not have any number part. Eg: .12345
3351                                 return $sign . $groupedNumber;
3352                         }
3353                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3354                         while ( $start > 0 ) {
3355                                 $match = $matches[0][$numMatches - 1];
3356                                 $matchLen = strlen( $match );
3357                                 $start = $end - $matchLen;
3358                                 if ( $start < 0 ) {
3359                                         $start = 0;
3360                                 }
3361                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3362                                 $end = $start;
3363                                 if ( $numMatches > 1 ) {
3364                                         // use the last pattern for the rest of the number
3365                                         $numMatches--;
3366                                 }
3367                                 if ( $start > 0 ) {
3368                                         $groupedNumber = "," . $groupedNumber;
3369                                 }
3370                         }
3371                         return $sign . $groupedNumber;
3372                 }
3373         }
3374
3375         /**
3376          * @return string
3377          */
3378         function digitGroupingPattern() {
3379                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3380         }
3381
3382         /**
3383          * @return array
3384          */
3385         function digitTransformTable() {
3386                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3387         }
3388
3389         /**
3390          * @return array
3391          */
3392         function separatorTransformTable() {
3393                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3394         }
3395
3396         /**
3397          * @return int|null
3398          */
3399         function minimumGroupingDigits() {
3400                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3401         }
3402
3403         /**
3404          * Take a list of strings and build a locale-friendly comma-separated
3405          * list, using the local comma-separator message.
3406          * The last two strings are chained with an "and".
3407          * NOTE: This function will only work with standard numeric array keys (0, 1, 2…)
3408          *
3409          * @param string[] $l
3410          * @return string
3411          */
3412         function listToText( array $l ) {
3413                 $m = count( $l ) - 1;
3414                 if ( $m < 0 ) {
3415                         return '';
3416                 }
3417                 if ( $m > 0 ) {
3418                         $and = $this->msg( 'and' )->escaped();
3419                         $space = $this->msg( 'word-separator' )->escaped();
3420                         if ( $m > 1 ) {
3421                                 $comma = $this->msg( 'comma-separator' )->escaped();
3422                         }
3423                 }
3424                 $s = $l[$m];
3425                 for ( $i = $m - 1; $i >= 0; $i-- ) {
3426                         if ( $i == $m - 1 ) {
3427                                 $s = $l[$i] . $and . $space . $s;
3428                         } else {
3429                                 $s = $l[$i] . $comma . $s;
3430                         }
3431                 }
3432                 return $s;
3433         }
3434
3435         /**
3436          * Take a list of strings and build a locale-friendly comma-separated
3437          * list, using the local comma-separator message.
3438          * @param string[] $list Array of strings to put in a comma list
3439          * @return string
3440          */
3441         function commaList( array $list ) {
3442                 return implode(
3443                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3444                         $list
3445                 );
3446         }
3447
3448         /**
3449          * Take a list of strings and build a locale-friendly semicolon-separated
3450          * list, using the local semicolon-separator message.
3451          * @param string[] $list Array of strings to put in a semicolon list
3452          * @return string
3453          */
3454         function semicolonList( array $list ) {
3455                 return implode(
3456                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3457                         $list
3458                 );
3459         }
3460
3461         /**
3462          * Same as commaList, but separate it with the pipe instead.
3463          * @param string[] $list Array of strings to put in a pipe list
3464          * @return string
3465          */
3466         function pipeList( array $list ) {
3467                 return implode(
3468                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3469                         $list
3470                 );
3471         }
3472
3473         /**
3474          * This method is deprecated since 1.31 and kept as alias for truncateForDatabase, which
3475          * has replaced it. This method provides truncation suitable for DB.
3476          *
3477          * The database offers limited byte lengths for some columns in the database;
3478          * multi-byte character sets mean we need to ensure that only whole characters
3479          * are included, otherwise broken characters can be passed to the user.
3480          *
3481          * @deprecated since 1.31, use truncateForDatabase or truncateForVisual as appropriate.
3482          *
3483          * @param string $string String to truncate
3484          * @param int $length Maximum length (including ellipsis)
3485          * @param string $ellipsis String to append to the truncated text
3486          * @param bool $adjustLength Subtract length of ellipsis from $length.
3487          *      $adjustLength was introduced in 1.18, before that behaved as if false.
3488          * @return string
3489          */
3490         function truncate( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3491                 return $this->truncateForDatabase( $string, $length, $ellipsis, $adjustLength );
3492         }
3493
3494         /**
3495          * Truncate a string to a specified length in bytes, appending an optional
3496          * string (e.g. for ellipsis)
3497          *
3498          * If $length is negative, the string will be truncated from the beginning
3499          *
3500          * @since 1.31
3501          *
3502          * @param string $string String to truncate
3503          * @param int $length Maximum length in bytes
3504          * @param string $ellipsis String to append to the end of truncated text
3505          * @param bool $adjustLength Subtract length of ellipsis from $length
3506          *
3507          * @return string
3508          */
3509         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3510                 return $this->truncateInternal(
3511                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3512                 );
3513         }
3514
3515         /**
3516          * Truncate a string to a specified number of characters, appending an optional
3517          * string (e.g. for ellipsis).
3518          *
3519          * This provides multibyte version of truncate() method of this class, suitable for truncation
3520          * based on number of characters, instead of number of bytes.
3521          *
3522          * If $length is negative, the string will be truncated from the beginning.
3523          *
3524          * @since 1.31
3525          *
3526          * @param string $string String to truncate
3527          * @param int $length Maximum number of characters
3528          * @param string $ellipsis String to append to the end of truncated text
3529          * @param bool $adjustLength Subtract length of ellipsis from $length
3530          *
3531          * @return string
3532          */
3533         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3534                 // Passing encoding to mb_strlen and mb_substr is optional.
3535                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3536                 // explicit specification of encoding is skipped.
3537                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3538                 return $this->truncateInternal(
3539                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3540                 );
3541         }
3542
3543         /**
3544          * Internal method used for truncation. This method abstracts text truncation into
3545          * one common method, allowing users to provide length measurement function and
3546          * function for finding substring.
3547          *
3548          * For usages, see truncateForDatabase and truncateForVisual.
3549          *
3550          * @param string $string String to truncate
3551          * @param int $length Maximum length of final text
3552          * @param string $ellipsis String to append to the end of truncated text
3553          * @param bool $adjustLength Subtract length of ellipsis from $length
3554          * @param callable $measureLength Callable function used for determining the length of text
3555          * @param callable $getSubstring Callable function used for getting the substrings
3556          *
3557          * @return string
3558          */
3559         private function truncateInternal(
3560                 $string, $length, $ellipsis, $adjustLength, $measureLength, $getSubstring
3561         ) {
3562                 if ( !is_callable( $measureLength ) || !is_callable( $getSubstring ) ) {
3563                         throw new InvalidArgumentException( 'Invalid callback provided' );
3564                 }
3565
3566                 # Check if there is no need to truncate
3567                 if ( $measureLength( $string ) <= abs( $length ) ) {
3568                         return $string; // no need to truncate
3569                 }
3570
3571                 # Use the localized ellipsis character
3572                 if ( $ellipsis == '...' ) {
3573                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3574                 }
3575                 if ( $length == 0 ) {
3576                         return $ellipsis; // convention
3577                 }
3578
3579                 $stringOriginal = $string;
3580                 # If ellipsis length is >= $length then we can't apply $adjustLength
3581                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3582                         $string = $ellipsis; // this can be slightly unexpected
3583                 # Otherwise, truncate and add ellipsis...
3584                 } else {
3585                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3586                         if ( $length > 0 ) {
3587                                 $length -= $ellipsisLength;
3588                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3589                                 $string = $this->removeBadCharLast( $string );
3590                                 $string = rtrim( $string );
3591                                 $string = $string . $ellipsis;
3592                         } else {
3593                                 $length += $ellipsisLength;
3594                                 $string = $getSubstring( $string, $length ); // ...xyz
3595                                 $string = $this->removeBadCharFirst( $string );
3596                                 $string = ltrim( $string );
3597                                 $string = $ellipsis . $string;
3598                         }
3599                 }
3600
3601                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3602                 # This check is *not* redundant if $adjustLength, due to the single case where
3603                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3604                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3605                         return $string;
3606                 } else {
3607                         return $stringOriginal;
3608                 }
3609         }
3610
3611         /**
3612          * Remove bytes that represent an incomplete Unicode character
3613          * at the end of string (e.g. bytes of the char are missing)
3614          *
3615          * @param string $string
3616          * @return string
3617          */
3618         protected function removeBadCharLast( $string ) {
3619                 if ( $string != '' ) {
3620                         $char = ord( $string[strlen( $string ) - 1] );
3621                         $m = [];
3622                         if ( $char >= 0xc0 ) {
3623                                 # We got the first byte only of a multibyte char; remove it.
3624                                 $string = substr( $string, 0, -1 );
3625                         } elseif ( $char >= 0x80 &&
3626                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3627                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3628                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3629                         ) {
3630                                 # We chopped in the middle of a character; remove it
3631                                 $string = $m[1];
3632                         }
3633                 }
3634                 return $string;
3635         }
3636
3637         /**
3638          * Remove bytes that represent an incomplete Unicode character
3639          * at the start of string (e.g. bytes of the char are missing)
3640          *
3641          * @param string $string
3642          * @return string
3643          */
3644         protected function removeBadCharFirst( $string ) {
3645                 if ( $string != '' ) {
3646                         $char = ord( $string[0] );
3647                         if ( $char >= 0x80 && $char < 0xc0 ) {
3648                                 # We chopped in the middle of a character; remove the whole thing
3649                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3650                         }
3651                 }
3652                 return $string;
3653         }
3654
3655         /**
3656          * Truncate a string of valid HTML to a specified length in bytes,
3657          * appending an optional string (e.g. for ellipses), and return valid HTML
3658          *
3659          * This is only intended for styled/linked text, such as HTML with
3660          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3661          * Also, this will not detect things like "display:none" CSS.
3662          *
3663          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3664          *
3665          * @param string $text HTML string to truncate
3666          * @param int $length (zero/positive) Maximum length (including ellipses)
3667          * @param string $ellipsis String to append to the truncated text
3668          * @return string
3669          */
3670         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3671                 # Use the localized ellipsis character
3672                 if ( $ellipsis == '...' ) {
3673                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3674                 }
3675                 # Check if there is clearly no need to truncate
3676                 if ( $length <= 0 ) {
3677                         return $ellipsis; // no text shown, nothing to format (convention)
3678                 } elseif ( strlen( $text ) <= $length ) {
3679                         return $text; // string short enough even *with* HTML (short-circuit)
3680                 }
3681
3682                 $dispLen = 0; // innerHTML legth so far
3683                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3684                 $tagType = 0; // 0-open, 1-close
3685                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3686                 $entityState = 0; // 0-not entity, 1-entity
3687                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3688                 $openTags = []; // open tag stack
3689                 $maybeState = null; // possible truncation state
3690
3691                 $textLen = strlen( $text );
3692                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3693                 for ( $pos = 0; true; ++$pos ) {
3694                         # Consider truncation once the display length has reached the maximim.
3695                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3696                         # Check that we're not in the middle of a bracket/entity...
3697                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3698                                 if ( !$testingEllipsis ) {
3699                                         $testingEllipsis = true;
3700                                         # Save where we are; we will truncate here unless there turn out to
3701                                         # be so few remaining characters that truncation is not necessary.
3702                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3703                                                 $maybeState = [ $ret, $openTags ]; // save state
3704                                         }
3705                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3706                                         # String in fact does need truncation, the truncation point was OK.
3707                                         list( $ret, $openTags ) = $maybeState; // reload state
3708                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3709                                         $ret .= $ellipsis; // add ellipsis
3710                                         break;
3711                                 }
3712                         }
3713                         if ( $pos >= $textLen ) {
3714                                 break; // extra iteration just for above checks
3715                         }
3716
3717                         # Read the next char...
3718                         $ch = $text[$pos];
3719                         $lastCh = $pos ? $text[$pos - 1] : '';
3720                         $ret .= $ch; // add to result string
3721                         if ( $ch == '<' ) {
3722                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3723                                 $entityState = 0; // for bad HTML
3724                                 $bracketState = 1; // tag started (checking for backslash)
3725                         } elseif ( $ch == '>' ) {
3726                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3727                                 $entityState = 0; // for bad HTML
3728                                 $bracketState = 0; // out of brackets
3729                         } elseif ( $bracketState == 1 ) {
3730                                 if ( $ch == '/' ) {
3731                                         $tagType = 1; // close tag (e.g. "</span>")
3732                                 } else {
3733                                         $tagType = 0; // open tag (e.g. "<span>")
3734                                         $tag .= $ch;
3735                                 }
3736                                 $bracketState = 2; // building tag name
3737                         } elseif ( $bracketState == 2 ) {
3738                                 if ( $ch != ' ' ) {
3739                                         $tag .= $ch;
3740                                 } else {
3741                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3742                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3743                                 }
3744                         } elseif ( $bracketState == 0 ) {
3745                                 if ( $entityState ) {
3746                                         if ( $ch == ';' ) {
3747                                                 $entityState = 0;
3748                                                 $dispLen++; // entity is one displayed char
3749                                         }
3750                                 } else {
3751                                         if ( $neLength == 0 && !$maybeState ) {
3752                                                 // Save state without $ch. We want to *hit* the first
3753                                                 // display char (to get tags) but not *use* it if truncating.
3754                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3755                                         }
3756                                         if ( $ch == '&' ) {
3757                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3758                                         } else {
3759                                                 $dispLen++; // this char is displayed
3760                                                 // Add the next $max display text chars after this in one swoop...
3761                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3762                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3763                                                 $dispLen += $skipped;
3764                                                 $pos += $skipped;
3765                                         }
3766                                 }
3767                         }
3768                 }
3769                 // Close the last tag if left unclosed by bad HTML
3770                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3771                 while ( count( $openTags ) > 0 ) {
3772                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3773                 }
3774                 return $ret;
3775         }
3776
3777         /**
3778          * truncateHtml() helper function
3779          * like strcspn() but adds the skipped chars to $ret
3780          *
3781          * @param string $ret
3782          * @param string $text
3783          * @param string $search
3784          * @param int $start
3785          * @param null|int $len
3786          * @return int
3787          */
3788         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3789                 if ( $len === null ) {
3790                         $len = -1; // -1 means "no limit" for strcspn
3791                 } elseif ( $len < 0 ) {
3792                         $len = 0; // sanity
3793                 }
3794                 $skipCount = 0;
3795                 if ( $start < strlen( $text ) ) {
3796                         $skipCount = strcspn( $text, $search, $start, $len );
3797                         $ret .= substr( $text, $start, $skipCount );
3798                 }
3799                 return $skipCount;
3800         }
3801
3802         /**
3803          * truncateHtml() helper function
3804          * (a) push or pop $tag from $openTags as needed
3805          * (b) clear $tag value
3806          * @param string &$tag Current HTML tag name we are looking at
3807          * @param int $tagType (0-open tag, 1-close tag)
3808          * @param string $lastCh Character before the '>' that ended this tag
3809          * @param array &$openTags Open tag stack (not accounting for $tag)
3810          */
3811         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3812                 $tag = ltrim( $tag );
3813                 if ( $tag != '' ) {
3814                         if ( $tagType == 0 && $lastCh != '/' ) {
3815                                 $openTags[] = $tag; // tag opened (didn't close itself)
3816                         } elseif ( $tagType == 1 ) {
3817                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3818                                         array_pop( $openTags ); // tag closed
3819                                 }
3820                         }
3821                         $tag = '';
3822                 }
3823         }
3824
3825         /**
3826          * Grammatical transformations, needed for inflected languages
3827          * Invoked by putting {{grammar:case|word}} in a message
3828          *
3829          * @param string $word
3830          * @param string $case
3831          * @return string
3832          */
3833         function convertGrammar( $word, $case ) {
3834                 global $wgGrammarForms;
3835                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3836                         return $wgGrammarForms[$this->getCode()][$case][$word];
3837                 }
3838
3839                 $grammarTransformations = $this->getGrammarTransformations();
3840
3841                 if ( isset( $grammarTransformations[$case] ) ) {
3842                         $forms = $grammarTransformations[$case];
3843
3844                         // Some names of grammar rules are aliases for other rules.
3845                         // In such cases the value is a string rather than object,
3846                         // so load the actual rules.
3847                         if ( is_string( $forms ) ) {
3848                                 $forms = $grammarTransformations[$forms];
3849                         }
3850
3851                         foreach ( array_values( $forms ) as $rule ) {
3852                                 $form = $rule[0];
3853
3854                                 if ( $form === '@metadata' ) {
3855                                         continue;
3856                                 }
3857
3858                                 $replacement = $rule[1];
3859
3860                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3861                                 $patternMatches = preg_match( $regex, $word );
3862
3863                                 if ( $patternMatches === false ) {
3864                                         wfLogWarning(
3865                                                 'An error occurred while processing grammar. ' .
3866                                                 "Word: '$word'. Regex: /$form/."
3867                                         );
3868                                 } elseif ( $patternMatches === 1 ) {
3869                                         $word = preg_replace( $regex, $replacement, $word );
3870
3871                                         break;
3872                                 }
3873                         }
3874                 }
3875
3876                 return $word;
3877         }
3878
3879         /**
3880          * Get the grammar forms for the content language
3881          * @return array Array of grammar forms
3882          * @since 1.20
3883          */
3884         function getGrammarForms() {
3885                 global $wgGrammarForms;
3886                 if ( isset( $wgGrammarForms[$this->getCode()] )
3887                         && is_array( $wgGrammarForms[$this->getCode()] )
3888                 ) {
3889                         return $wgGrammarForms[$this->getCode()];
3890                 }
3891
3892                 return [];
3893         }
3894
3895         /**
3896          * Get the grammar transformations data for the language.
3897          * Used like grammar forms, with {{GRAMMAR}} and cases,
3898          * but uses pairs of regexes and replacements instead of code.
3899          *
3900          * @return array[] Array of grammar transformations.
3901          * @throws MWException
3902          * @since 1.28
3903          */
3904         public function getGrammarTransformations() {
3905                 $languageCode = $this->getCode();
3906
3907                 if ( self::$grammarTransformations === null ) {
3908                         self::$grammarTransformations = new MapCacheLRU( 10 );
3909                 }
3910
3911                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3912                         return self::$grammarTransformations->get( $languageCode );
3913                 }
3914
3915                 $data = [];
3916
3917                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3918                 if ( is_readable( $grammarDataFile ) ) {
3919                         $data = FormatJson::decode(
3920                                 file_get_contents( $grammarDataFile ),
3921                                 true
3922                         );
3923
3924                         if ( $data === null ) {
3925                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3926                         }
3927
3928                         self::$grammarTransformations->set( $languageCode, $data );
3929                 }
3930
3931                 return $data;
3932         }
3933
3934         /**
3935          * Provides an alternative text depending on specified gender.
3936          * Usage {{gender:username|masculine|feminine|unknown}}.
3937          * username is optional, in which case the gender of current user is used,
3938          * but only in (some) interface messages; otherwise default gender is used.
3939          *
3940          * If no forms are given, an empty string is returned. If only one form is
3941          * given, it will be returned unconditionally. These details are implied by
3942          * the caller and cannot be overridden in subclasses.
3943          *
3944          * If three forms are given, the default is to use the third (unknown) form.
3945          * If fewer than three forms are given, the default is to use the first (masculine) form.
3946          * These details can be overridden in subclasses.
3947          *
3948          * @param string $gender
3949          * @param array $forms
3950          *
3951          * @return string
3952          */
3953         function gender( $gender, $forms ) {
3954                 if ( !count( $forms ) ) {
3955                         return '';
3956                 }
3957                 $forms = $this->preConvertPlural( $forms, 2 );
3958                 if ( $gender === 'male' ) {
3959                         return $forms[0];
3960                 }
3961                 if ( $gender === 'female' ) {
3962                         return $forms[1];
3963                 }
3964                 return $forms[2] ?? $forms[0];
3965         }
3966
3967         /**
3968          * Plural form transformations, needed for some languages.
3969          * For example, there are 3 form of plural in Russian and Polish,
3970          * depending on "count mod 10". See [[w:Plural]]
3971          * For English it is pretty simple.
3972          *
3973          * Invoked by putting {{plural:count|wordform1|wordform2}}
3974          * or {{plural:count|wordform1|wordform2|wordform3}}
3975          *
3976          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3977          *
3978          * @param int $count Non-localized number
3979          * @param array $forms Different plural forms
3980          * @return string Correct form of plural for $count in this language
3981          */
3982         function convertPlural( $count, $forms ) {
3983                 // Handle explicit n=pluralform cases
3984                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3985                 if ( is_string( $forms ) ) {
3986                         return $forms;
3987                 }
3988                 if ( !count( $forms ) ) {
3989                         return '';
3990                 }
3991
3992                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3993                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3994                 return $forms[$pluralForm];
3995         }
3996
3997         /**
3998          * Handles explicit plural forms for Language::convertPlural()
3999          *
4000          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
4001          * If an explicitly defined plural form matches the $count, then
4002          * string value returned, otherwise array returned for further consideration
4003          * by CLDR rules or overridden convertPlural().
4004          *
4005          * @since 1.23
4006          *
4007          * @param int $count Non-localized number
4008          * @param array $forms Different plural forms
4009          *
4010          * @return array|string
4011          */
4012         protected function handleExplicitPluralForms( $count, array $forms ) {
4013                 foreach ( $forms as $index => $form ) {
4014                         if ( preg_match( '/\d+=/i', $form ) ) {
4015                                 $pos = strpos( $form, '=' );
4016                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4017                                         return substr( $form, $pos + 1 );
4018                                 }
4019                                 unset( $forms[$index] );
4020                         }
4021                 }
4022                 return array_values( $forms );
4023         }
4024
4025         /**
4026          * Checks that convertPlural was given an array and pads it to requested
4027          * amount of forms by copying the last one.
4028          *
4029          * @param array $forms Array of forms given to convertPlural
4030          * @param int $count How many forms should there be at least
4031          * @return array Padded array of forms or an exception if not an array
4032          */
4033         protected function preConvertPlural( /* Array */ $forms, $count ) {
4034                 while ( count( $forms ) < $count ) {
4035                         $forms[] = $forms[count( $forms ) - 1];
4036                 }
4037                 return $forms;
4038         }
4039
4040         /**
4041          * Wraps argument with unicode control characters for directionality safety
4042          *
4043          * This solves the problem where directionality-neutral characters at the edge of
4044          * the argument string get interpreted with the wrong directionality from the
4045          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4046          *
4047          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4048          * directionality of the argument string, using the BIDI algorithm's own "First
4049          * strong directional codepoint" rule. Essentially, this works round the fact that
4050          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4051          * direction inference). The latter is cleaner but still not widely supported.
4052          *
4053          * @param string $text Text to wrap
4054          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4055          */
4056         public function embedBidi( $text = '' ) {
4057                 $dir = self::strongDirFromContent( $text );
4058                 if ( $dir === 'ltr' ) {
4059                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4060                         return self::$lre . $text . self::$pdf;
4061                 }
4062                 if ( $dir === 'rtl' ) {
4063                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4064                         return self::$rle . $text . self::$pdf;
4065                 }
4066                 // No strong directionality: do not wrap
4067                 return $text;
4068         }
4069
4070         /**
4071          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4072          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4073          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4074          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4075          * match up with it.
4076          *
4077          * @param string $str The validated block duration in English
4078          * @param User $user User object to use timezone from or null for $wgUser
4079          * @param int $now Current timestamp, for formatting relative block durations
4080          * @return string Somehow translated block duration
4081          * @see LanguageFi.php for example implementation
4082          */
4083         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4084                 $duration = SpecialBlock::getSuggestedDurations( $this );
4085                 foreach ( $duration as $show => $value ) {
4086                         if ( strcmp( $str, $value ) == 0 ) {
4087                                 return htmlspecialchars( trim( $show ) );
4088                         }
4089                 }
4090
4091                 if ( wfIsInfinity( $str ) ) {
4092                         foreach ( $duration as $show => $value ) {
4093                                 if ( wfIsInfinity( $value ) ) {
4094                                         return htmlspecialchars( trim( $show ) );
4095                                 }
4096                         }
4097                 }
4098
4099                 // If all else fails, return a standard duration or timestamp description.
4100                 $time = strtotime( $str, $now );
4101                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4102                         return $str;
4103                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4104                         // The result differs based on current time, so the difference
4105                         // is a fixed duration length.
4106                         return $this->formatDuration( $time - $now );
4107                 } else { // It's an absolute timestamp.
4108                         if ( $time === 0 ) {
4109                                 // wfTimestamp() handles 0 as current time instead of epoch.
4110                                 $time = '19700101000000';
4111                         }
4112                         if ( $user ) {
4113                                 return $this->userTimeAndDate( $time, $user );
4114                         }
4115                         return $this->timeanddate( $time );
4116                 }
4117         }
4118
4119         /**
4120          * languages like Chinese need to be segmented in order for the diff
4121          * to be of any use
4122          *
4123          * @param string $text
4124          * @return string
4125          */
4126         public function segmentForDiff( $text ) {
4127                 return $text;
4128         }
4129
4130         /**
4131          * and unsegment to show the result
4132          *
4133          * @param string $text
4134          * @return string
4135          */
4136         public function unsegmentForDiff( $text ) {
4137                 return $text;
4138         }
4139
4140         /**
4141          * Return the LanguageConverter used in the Language
4142          *
4143          * @since 1.19
4144          * @return LanguageConverter
4145          */
4146         public function getConverter() {
4147                 return $this->mConverter;
4148         }
4149
4150         /**
4151          * convert text to a variant
4152          *
4153          * @param string $text text to convert
4154          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4155          *      variant (if logged in), or the project default variant
4156          * @return string the converted string
4157          */
4158         public function autoConvert( $text, $variant = false ) {
4159                 return $this->mConverter->autoConvert( $text, $variant );
4160         }
4161
4162         /**
4163          * convert text to all supported variants
4164          *
4165          * @param string $text
4166          * @return array
4167          */
4168         public function autoConvertToAllVariants( $text ) {
4169                 return $this->mConverter->autoConvertToAllVariants( $text );
4170         }
4171
4172         /**
4173          * convert text to different variants of a language.
4174          *
4175          * @param string $text
4176          * @return string
4177          */
4178         public function convert( $text ) {
4179                 return $this->mConverter->convert( $text );
4180         }
4181
4182         /**
4183          * Convert a Title object to a string in the preferred variant
4184          *
4185          * @param Title $title
4186          * @return string
4187          */
4188         public function convertTitle( $title ) {
4189                 return $this->mConverter->convertTitle( $title );
4190         }
4191
4192         /**
4193          * Convert a namespace index to a string in the preferred variant
4194          *
4195          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4196          * @param string|null $variant variant to convert to, or null to use the user's preferred
4197          *      variant (if logged in), or the project default variant
4198          * @return string a string representation of the namespace
4199          */
4200         public function convertNamespace( $ns, $variant = null ) {
4201                 return $this->mConverter->convertNamespace( $ns, $variant );
4202         }
4203
4204         /**
4205          * Check if this is a language with variants
4206          *
4207          * @return bool
4208          */
4209         public function hasVariants() {
4210                 return count( $this->getVariants() ) > 1;
4211         }
4212
4213         /**
4214          * Check if the language has the specific variant
4215          *
4216          * @since 1.19
4217          * @param string $variant
4218          * @return bool
4219          */
4220         public function hasVariant( $variant ) {
4221                 return (bool)$this->mConverter->validateVariant( $variant );
4222         }
4223
4224         /**
4225          * Perform output conversion on a string, and encode for safe HTML output.
4226          * @param string $text Text to be converted
4227          * @param bool $isTitle Whether this conversion is for the article title
4228          * @return string
4229          * @todo this should get integrated somewhere sane
4230          */
4231         public function convertHtml( $text, $isTitle = false ) {
4232                 return htmlspecialchars( $this->convert( $text, $isTitle ) );
4233         }
4234
4235         /**
4236          * @param string $key
4237          * @return string
4238          */
4239         public function convertCategoryKey( $key ) {
4240                 return $this->mConverter->convertCategoryKey( $key );
4241         }
4242
4243         /**
4244          * Get the list of variants supported by this language
4245          * see sample implementation in LanguageZh.php
4246          *
4247          * @return string[] An array of language codes
4248          */
4249         public function getVariants() {
4250                 return $this->mConverter->getVariants();
4251         }
4252
4253         /**
4254          * @return string
4255          */
4256         public function getPreferredVariant() {
4257                 return $this->mConverter->getPreferredVariant();
4258         }
4259
4260         /**
4261          * @return string
4262          */
4263         public function getDefaultVariant() {
4264                 return $this->mConverter->getDefaultVariant();
4265         }
4266
4267         /**
4268          * @return string
4269          */
4270         public function getURLVariant() {
4271                 return $this->mConverter->getURLVariant();
4272         }
4273
4274         /**
4275          * If a language supports multiple variants, it is
4276          * possible that non-existing link in one variant
4277          * actually exists in another variant. this function
4278          * tries to find it. See e.g. LanguageZh.php
4279          * The input parameters may be modified upon return
4280          *
4281          * @param string &$link The name of the link
4282          * @param Title &$nt The title object of the link
4283          * @param bool $ignoreOtherCond To disable other conditions when
4284          *   we need to transclude a template or update a category's link
4285          */
4286         public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
4287                 $this->mConverter->findVariantLink( $link, $nt, $ignoreOtherCond );
4288         }
4289
4290         /**
4291          * returns language specific options used by User::getPageRenderHash()
4292          * for example, the preferred language variant
4293          *
4294          * @return string
4295          */
4296         function getExtraHashOptions() {
4297                 return $this->mConverter->getExtraHashOptions();
4298         }
4299
4300         /**
4301          * For languages that support multiple variants, the title of an
4302          * article may be displayed differently in different variants. this
4303          * function returns the apporiate title defined in the body of the article.
4304          *
4305          * @return string
4306          */
4307         public function getParsedTitle() {
4308                 return $this->mConverter->getParsedTitle();
4309         }
4310
4311         /**
4312          * Refresh the cache of conversion tables when
4313          * MediaWiki:Conversiontable* is updated.
4314          *
4315          * @param Title $title The Title of the page being updated
4316          */
4317         public function updateConversionTable( Title $title ) {
4318                 $this->mConverter->updateConversionTable( $title );
4319         }
4320
4321         /**
4322          * Prepare external link text for conversion. When the text is
4323          * a URL, it shouldn't be converted, and it'll be wrapped in
4324          * the "raw" tag (-{R| }-) to prevent conversion.
4325          *
4326          * This function is called "markNoConversion" for historical
4327          * reasons *BUT DIFFERS SIGNIFICANTLY* from
4328          * LanguageConverter::markNoConversion(), with which it is easily
4329          * confused.
4330          *
4331          * @param string $text Text to be used for external link
4332          * @param bool $noParse Wrap it without confirming it's a real URL first
4333          * @return string The tagged text
4334          * @deprecated since 1.32, use LanguageConverter::markNoConversion()
4335          *  instead.
4336          */
4337         public function markNoConversion( $text, $noParse = false ) {
4338                 wfDeprecated( __METHOD__, '1.32' );
4339                 // Excluding protocal-relative URLs may avoid many false positives.
4340                 if ( $noParse || preg_match( '/^(?:' . wfUrlProtocolsWithoutProtRel() . ')/', $text ) ) {
4341                         return $this->mConverter->markNoConversion( $text );
4342                 } else {
4343                         return $text;
4344                 }
4345         }
4346
4347         /**
4348          * A regular expression to match legal word-trailing characters
4349          * which should be merged onto a link of the form [[foo]]bar.
4350          *
4351          * @return string
4352          */
4353         public function linkTrail() {
4354                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4355         }
4356
4357         /**
4358          * A regular expression character set to match legal word-prefixing
4359          * characters which should be merged onto a link of the form foo[[bar]].
4360          *
4361          * @return string
4362          */
4363         public function linkPrefixCharset() {
4364                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4365         }
4366
4367         /**
4368          * Get the "parent" language which has a converter to convert a "compatible" language
4369          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4370          *
4371          * @return Language|null
4372          * @since 1.22
4373          */
4374         public function getParentLanguage() {
4375                 if ( $this->mParentLanguage !== false ) {
4376                         return $this->mParentLanguage;
4377                 }
4378
4379                 $code = explode( '-', $this->getCode() )[0];
4380                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4381                         $this->mParentLanguage = null;
4382                         return null;
4383                 }
4384                 $lang = self::factory( $code );
4385                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4386                         $this->mParentLanguage = null;
4387                         return null;
4388                 }
4389
4390                 $this->mParentLanguage = $lang;
4391                 return $lang;
4392         }
4393
4394         /**
4395          * Compare with an other language object
4396          *
4397          * @since 1.28
4398          * @param Language $lang
4399          * @return bool
4400          */
4401         public function equals( Language $lang ) {
4402                 return $lang->getCode() === $this->mCode;
4403         }
4404
4405         /**
4406          * Get the internal language code for this language object
4407          *
4408          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4409          * htmlspecialchars() or similar
4410          *
4411          * @return string
4412          */
4413         public function getCode() {
4414                 return $this->mCode;
4415         }
4416
4417         /**
4418          * Get the code in BCP 47 format which we can use
4419          * inside of html lang="" tags.
4420          *
4421          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4422          * htmlspecialchars() or similar.
4423          *
4424          * @since 1.19
4425          * @return string
4426          */
4427         public function getHtmlCode() {
4428                 if ( is_null( $this->mHtmlCode ) ) {
4429                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4430                 }
4431                 return $this->mHtmlCode;
4432         }
4433
4434         /**
4435          * @param string $code
4436          */
4437         public function setCode( $code ) {
4438                 $this->mCode = $code;
4439                 // Ensure we don't leave incorrect cached data lying around
4440                 $this->mHtmlCode = null;
4441                 $this->mParentLanguage = false;
4442         }
4443
4444         /**
4445          * Get the language code from a file name. Inverse of getFileName()
4446          * @param string $filename $prefix . $languageCode . $suffix
4447          * @param string $prefix Prefix before the language code
4448          * @param string $suffix Suffix after the language code
4449          * @return string Language code, or false if $prefix or $suffix isn't found
4450          */
4451         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4452                 $m = null;
4453                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4454                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4455                 if ( !count( $m ) ) {
4456                         return false;
4457                 }
4458                 return str_replace( '_', '-', strtolower( $m[1] ) );
4459         }
4460
4461         /**
4462          * @param string $code
4463          * @param bool $fallback Whether we're going through language fallback chain
4464          * @return string Name of the language class
4465          */
4466         public static function classFromCode( $code, $fallback = true ) {
4467                 if ( $fallback && $code == 'en' ) {
4468                         return 'Language';
4469                 } else {
4470                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4471                 }
4472         }
4473
4474         /**
4475          * Get the name of a file for a certain language code
4476          * @param string $prefix Prepend this to the filename
4477          * @param string $code Language code
4478          * @param string $suffix Append this to the filename
4479          * @throws MWException
4480          * @return string $prefix . $mangledCode . $suffix
4481          */
4482         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4483                 if ( !self::isValidBuiltInCode( $code ) ) {
4484                         throw new MWException( "Invalid language code \"$code\"" );
4485                 }
4486
4487                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4488         }
4489
4490         /**
4491          * @param string $code
4492          * @return string
4493          */
4494         public static function getMessagesFileName( $code ) {
4495                 global $IP;
4496                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4497                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4498                 return $file;
4499         }
4500
4501         /**
4502          * @param string $code
4503          * @return string
4504          * @throws MWException
4505          * @since 1.23
4506          */
4507         public static function getJsonMessagesFileName( $code ) {
4508                 global $IP;
4509
4510                 if ( !self::isValidBuiltInCode( $code ) ) {
4511                         throw new MWException( "Invalid language code \"$code\"" );
4512                 }
4513
4514                 return "$IP/languages/i18n/$code.json";
4515         }
4516
4517         /**
4518          * Get the first fallback for a given language.
4519          *
4520          * @param string $code
4521          *
4522          * @return bool|string
4523          */
4524         public static function getFallbackFor( $code ) {
4525                 $fallbacks = self::getFallbacksFor( $code );
4526                 if ( $fallbacks ) {
4527                         return $fallbacks[0];
4528                 }
4529                 return false;
4530         }
4531
4532         /**
4533          * Get the ordered list of fallback languages.
4534          *
4535          * @since 1.19
4536          * @param string $code Language code
4537          * @return array Non-empty array, ending in "en"
4538          */
4539         public static function getFallbacksFor( $code ) {
4540                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4541                         return [];
4542                 }
4543                 // For unknown languages, fallbackSequence returns an empty array,
4544                 // hardcode fallback to 'en' in that case.
4545                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4546         }
4547
4548         /**
4549          * Get the ordered list of fallback languages, ending with the fallback
4550          * language chain for the site language.
4551          *
4552          * @since 1.22
4553          * @param string $code Language code
4554          * @return array Array( fallbacks, site fallbacks )
4555          */
4556         public static function getFallbacksIncludingSiteLanguage( $code ) {
4557                 global $wgLanguageCode;
4558
4559                 // Usually, we will only store a tiny number of fallback chains, so we
4560                 // keep them in static memory.
4561                 $cacheKey = "{$code}-{$wgLanguageCode}";
4562
4563                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4564                         $fallbacks = self::getFallbacksFor( $code );
4565
4566                         // Append the site's fallback chain, including the site language itself
4567                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4568                         array_unshift( $siteFallbacks, $wgLanguageCode );
4569
4570                         // Eliminate any languages already included in the chain
4571                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4572
4573                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4574                 }
4575                 return self::$fallbackLanguageCache[$cacheKey];
4576         }
4577
4578         /**
4579          * Get all messages for a given language
4580          * WARNING: this may take a long time. If you just need all message *keys*
4581          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4582          *
4583          * @param string $code
4584          *
4585          * @return array
4586          */
4587         public static function getMessagesFor( $code ) {
4588                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4589         }
4590
4591         /**
4592          * Get a message for a given language
4593          *
4594          * @param string $key
4595          * @param string $code
4596          *
4597          * @return string
4598          */
4599         public static function getMessageFor( $key, $code ) {
4600                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4601         }
4602
4603         /**
4604          * Get all message keys for a given language. This is a faster alternative to
4605          * array_keys( Language::getMessagesFor( $code ) )
4606          *
4607          * @since 1.19
4608          * @param string $code Language code
4609          * @return array Array of message keys (strings)
4610          */
4611         public static function getMessageKeysFor( $code ) {
4612                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4613         }
4614
4615         /**
4616          * @param string $talk
4617          * @return mixed
4618          */
4619         function fixVariableInNamespace( $talk ) {
4620                 if ( strpos( $talk, '$1' ) === false ) {
4621                         return $talk;
4622                 }
4623
4624                 global $wgMetaNamespace;
4625                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4626
4627                 # Allow grammar transformations
4628                 # Allowing full message-style parsing would make simple requests
4629                 # such as action=raw much more expensive than they need to be.
4630                 # This will hopefully cover most cases.
4631                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4632                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4633                 return str_replace( ' ', '_', $talk );
4634         }
4635
4636         /**
4637          * @param string $m
4638          * @return string
4639          */
4640         function replaceGrammarInNamespace( $m ) {
4641                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4642         }
4643
4644         /**
4645          * Decode an expiry (block, protection, etc) which has come from the DB
4646          *
4647          * @param string $expiry Database expiry String
4648          * @param bool|int $format True to process using language functions, or TS_ constant
4649          *     to return the expiry in a given timestamp
4650          * @param string $infinity If $format is not true, use this string for infinite expiry
4651          * @return string
4652          * @since 1.18
4653          */
4654         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4655                 static $dbInfinity;
4656                 if ( $dbInfinity === null ) {
4657                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4658                 }
4659
4660                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4661                         return $format === true
4662                                 ? $this->getMessageFromDB( 'infiniteblock' )
4663                                 : $infinity;
4664                 } else {
4665                         return $format === true
4666                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4667                                 : wfTimestamp( $format, $expiry );
4668                 }
4669         }
4670
4671         /**
4672          * Formats a time given in seconds into a string representation of that time.
4673          *
4674          * @param int|float $seconds
4675          * @param array $format An optional argument that formats the returned string in different ways:
4676          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4677          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4678          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4679          *     and friends.
4680          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4681          *     or 'avoidminutes'.
4682          * @return string
4683          */
4684         function formatTimePeriod( $seconds, $format = [] ) {
4685                 if ( !is_array( $format ) ) {
4686                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4687                 }
4688                 if ( !isset( $format['avoid'] ) ) {
4689                         $format['avoid'] = false;
4690                 }
4691                 if ( !isset( $format['noabbrevs'] ) ) {
4692                         $format['noabbrevs'] = false;
4693                 }
4694                 $secondsMsg = wfMessage(
4695                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4696                 $minutesMsg = wfMessage(
4697                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4698                 $hoursMsg = wfMessage(
4699                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4700                 $daysMsg = wfMessage(
4701                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4702
4703                 if ( round( $seconds * 10 ) < 100 ) {
4704                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4705                         $s = $secondsMsg->params( $s )->text();
4706                 } elseif ( round( $seconds ) < 60 ) {
4707                         $s = $this->formatNum( round( $seconds ) );
4708                         $s = $secondsMsg->params( $s )->text();
4709                 } elseif ( round( $seconds ) < 3600 ) {
4710                         $minutes = floor( $seconds / 60 );
4711                         $secondsPart = round( fmod( $seconds, 60 ) );
4712                         if ( $secondsPart == 60 ) {
4713                                 $secondsPart = 0;
4714                                 $minutes++;
4715                         }
4716                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4717                         $s .= ' ';
4718                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4719                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4720                         $hours = floor( $seconds / 3600 );
4721                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4722                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4723                         if ( $secondsPart == 60 ) {
4724                                 $secondsPart = 0;
4725                                 $minutes++;
4726                         }
4727                         if ( $minutes == 60 ) {
4728                                 $minutes = 0;
4729                                 $hours++;
4730                         }
4731                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4732                         $s .= ' ';
4733                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4734                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4735                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4736                         }
4737                 } else {
4738                         $days = floor( $seconds / 86400 );
4739                         if ( $format['avoid'] === 'avoidminutes' ) {
4740                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4741                                 if ( $hours == 24 ) {
4742                                         $hours = 0;
4743                                         $days++;
4744                                 }
4745                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4746                                 $s .= ' ';
4747                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4748                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4749                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4750                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4751                                 if ( $minutes == 60 ) {
4752                                         $minutes = 0;
4753                                         $hours++;
4754                                 }
4755                                 if ( $hours == 24 ) {
4756                                         $hours = 0;
4757                                         $days++;
4758                                 }
4759                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4760                                 $s .= ' ';
4761                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4762                                 $s .= ' ';
4763                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4764                         } else {
4765                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4766                                 $s .= ' ';
4767                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4768                         }
4769                 }
4770                 return $s;
4771         }
4772
4773         /**
4774          * Format a bitrate for output, using an appropriate
4775          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4776          *   the magnitude in question.
4777          *
4778          * This use base 1000. For base 1024 use formatSize(), for another base
4779          * see formatComputingNumbers().
4780          *
4781          * @param int $bps
4782          * @return string
4783          */
4784         function formatBitrate( $bps ) {
4785                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4786         }
4787
4788         /**
4789          * @param int $size Size of the unit
4790          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4791          * @param string $messageKey Message key to be uesd
4792          * @return string
4793          */
4794         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4795                 if ( $size <= 0 ) {
4796                         return str_replace( '$1', $this->formatNum( $size ),
4797                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4798                         );
4799                 }
4800                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4801                 $index = 0;
4802
4803                 $maxIndex = count( $sizes ) - 1;
4804                 while ( $size >= $boundary && $index < $maxIndex ) {
4805                         $index++;
4806                         $size /= $boundary;
4807                 }
4808
4809                 // For small sizes no decimal places necessary
4810                 $round = 0;
4811                 if ( $index > 1 ) {
4812                         // For MB and bigger two decimal places are smarter
4813                         $round = 2;
4814                 }
4815                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4816
4817                 $size = round( $size, $round );
4818                 $text = $this->getMessageFromDB( $msg );
4819                 return str_replace( '$1', $this->formatNum( $size ), $text );
4820         }
4821
4822         /**
4823          * Format a size in bytes for output, using an appropriate
4824          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4825          *
4826          * This method use base 1024. For base 1000 use formatBitrate(), for
4827          * another base see formatComputingNumbers()
4828          *
4829          * @param int $size Size to format
4830          * @return string Plain text (not HTML)
4831          */
4832         function formatSize( $size ) {
4833                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4834         }
4835
4836         /**
4837          * Make a list item, used by various special pages
4838          *
4839          * @param string $page Page link
4840          * @param string $details HTML safe text between brackets
4841          * @param bool $oppositedm Add the direction mark opposite to your
4842          *   language, to display text properly
4843          * @return string HTML escaped
4844          */
4845         function specialList( $page, $details, $oppositedm = true ) {
4846                 if ( !$details ) {
4847                         return $page;
4848                 }
4849
4850                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4851                 return $page .
4852                         $dirmark .
4853                         $this->msg( 'word-separator' )->escaped() .
4854                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4855         }
4856
4857         /**
4858          * Generate (prev x| next x) (20|50|100...) type links for paging
4859          *
4860          * @param Title $title Title object to link
4861          * @param int $offset
4862          * @param int $limit
4863          * @param array $query Optional URL query parameter string
4864          * @param bool $atend Optional param for specified if this is the last page
4865          * @return string
4866          */
4867         public function viewPrevNext( Title $title, $offset, $limit,
4868                 array $query = [], $atend = false
4869         ) {
4870                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4871
4872                 # Make 'previous' link
4873                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4874                 if ( $offset > 0 ) {
4875                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4876                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4877                 } else {
4878                         $plink = htmlspecialchars( $prev );
4879                 }
4880
4881                 # Make 'next' link
4882                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4883                 if ( $atend ) {
4884                         $nlink = htmlspecialchars( $next );
4885                 } else {
4886                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4887                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4888                 }
4889
4890                 # Make links to set number of items per page
4891                 $numLinks = [];
4892                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4893                         $numLinks[] = $this->numLink( $title, $offset, $num,
4894                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4895                 }
4896
4897                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4898                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4899         }
4900
4901         /**
4902          * Helper function for viewPrevNext() that generates links
4903          *
4904          * @param Title $title Title object to link
4905          * @param int $offset
4906          * @param int $limit
4907          * @param array $query Extra query parameters
4908          * @param string $link Text to use for the link; will be escaped
4909          * @param string $tooltipMsg Name of the message to use as tooltip
4910          * @param string $class Value of the "class" attribute of the link
4911          * @return string HTML fragment
4912          */
4913         private function numLink( Title $title, $offset, $limit, array $query, $link,
4914                 $tooltipMsg, $class
4915         ) {
4916                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4917                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4918                         ->numParams( $limit )->text();
4919
4920                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4921                         'title' => $tooltip, 'class' => $class ], $link );
4922         }
4923
4924         /**
4925          * Get the conversion rule title, if any.
4926          *
4927          * @return string
4928          */
4929         public function getConvRuleTitle() {
4930                 return $this->mConverter->getConvRuleTitle();
4931         }
4932
4933         /**
4934          * Get the compiled plural rules for the language
4935          * @since 1.20
4936          * @return array Associative array with plural form, and plural rule as key-value pairs
4937          */
4938         public function getCompiledPluralRules() {
4939                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4940                 $fallbacks = self::getFallbacksFor( $this->mCode );
4941                 if ( !$pluralRules ) {
4942                         foreach ( $fallbacks as $fallbackCode ) {
4943                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4944                                 if ( $pluralRules ) {
4945                                         break;
4946                                 }
4947                         }
4948                 }
4949                 return $pluralRules;
4950         }
4951
4952         /**
4953          * Get the plural rules for the language
4954          * @since 1.20
4955          * @return array Associative array with plural form number and plural rule as key-value pairs
4956          */
4957         public function getPluralRules() {
4958                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4959                 $fallbacks = self::getFallbacksFor( $this->mCode );
4960                 if ( !$pluralRules ) {
4961                         foreach ( $fallbacks as $fallbackCode ) {
4962                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4963                                 if ( $pluralRules ) {
4964                                         break;
4965                                 }
4966                         }
4967                 }
4968                 return $pluralRules;
4969         }
4970
4971         /**
4972          * Get the plural rule types for the language
4973          * @since 1.22
4974          * @return array Associative array with plural form number and plural rule type as key-value pairs
4975          */
4976         public function getPluralRuleTypes() {
4977                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4978                 $fallbacks = self::getFallbacksFor( $this->mCode );
4979                 if ( !$pluralRuleTypes ) {
4980                         foreach ( $fallbacks as $fallbackCode ) {
4981                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4982                                 if ( $pluralRuleTypes ) {
4983                                         break;
4984                                 }
4985                         }
4986                 }
4987                 return $pluralRuleTypes;
4988         }
4989
4990         /**
4991          * Find the index number of the plural rule appropriate for the given number
4992          * @param int $number
4993          * @return int The index number of the plural rule
4994          */
4995         public function getPluralRuleIndexNumber( $number ) {
4996                 $pluralRules = $this->getCompiledPluralRules();
4997                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4998                 return $form;
4999         }
5000
5001         /**
5002          * Find the plural rule type appropriate for the given number
5003          * For example, if the language is set to Arabic, getPluralType(5) should
5004          * return 'few'.
5005          * @since 1.22
5006          * @param int $number
5007          * @return string The name of the plural rule type, e.g. one, two, few, many
5008          */
5009         public function getPluralRuleType( $number ) {
5010                 $index = $this->getPluralRuleIndexNumber( $number );
5011                 $pluralRuleTypes = $this->getPluralRuleTypes();
5012                 if ( isset( $pluralRuleTypes[$index] ) ) {
5013                         return $pluralRuleTypes[$index];
5014                 } else {
5015                         return 'other';
5016                 }
5017         }
5018 }