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 use Wikimedia\Assert\Assert;
  31
  32 /**
  33  * Internationalisation code
  34  * @ingroup Language
  35  */
  36 class Language {
  37         /**
  38          * Return autonyms in fetchLanguageName(s).
  39          * @since 1.32
  40          */
  41         const AS_AUTONYMS = null;
  42
  43         /**
  44          * Return all known languages in fetchLanguageName(s).
  45          * @since 1.32
  46          */
  47         const ALL = 'all';
  48
  49         /**
  50          * Return in fetchLanguageName(s) only the languages for which we have at
  51          * least some localisation.
  52          * @since 1.32
  53          */
  54         const SUPPORTED = 'mwfile';
  55
  56         /**
  57          * @var LanguageConverter
  58          */
  59         public $mConverter;
  60
  61         public $mVariants, $mCode, $mLoaded = false;
  62         public $mMagicExtensions = [];
  63         private $mHtmlCode = null, $mParentLanguage = false;
  64
  65         public $dateFormatStrings = [];
  66         public $mExtendedSpecialPageAliases;
  67
  68         /** @var array|null */
  69         protected $namespaceNames;
  70         protected $mNamespaceIds, $namespaceAliases;
  71
  72         /**
  73          * ReplacementArray object caches
  74          */
  75         public $transformData = [];
  76
  77         /**
  78          * @var LocalisationCache
  79          */
  80         public static $dataCache;
  81
  82         public static $mLangObjCache = [];
  83
  84         /**
  85          * Return a fallback chain for messages in getFallbacksFor
  86          * @since 1.32
  87          */
  88         const MESSAGES_FALLBACKS = 0;
  89
  90         /**
  91          * Return a strict fallback chain in getFallbacksFor
  92          * @since 1.32
  93          */
  94         const STRICT_FALLBACKS = 1;
  95
  96         public static $mWeekdayMsgs = [
  97                 'sunday', 'monday', 'tuesday', 'wednesday', 'thursday',
  98                 'friday', 'saturday'
  99         ];
 100
 101         public static $mWeekdayAbbrevMsgs = [
 102                 'sun', 'mon', 'tue', 'wed', 'thu', 'fri', 'sat'
 103         ];
 104
 105         public static $mMonthMsgs = [
 106                 'january', 'february', 'march', 'april', 'may_long', 'june',
 107                 'july', 'august', 'september', 'october', 'november',
 108                 'december'
 109         ];
 110         public static $mMonthGenMsgs = [
 111                 'january-gen', 'february-gen', 'march-gen', 'april-gen', 'may-gen', 'june-gen',
 112                 'july-gen', 'august-gen', 'september-gen', 'october-gen', 'november-gen',
 113                 'december-gen'
 114         ];
 115         public static $mMonthAbbrevMsgs = [
 116                 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug',
 117                 'sep', 'oct', 'nov', 'dec'
 118         ];
 119
 120         public static $mIranianCalendarMonthMsgs = [
 121                 'iranian-calendar-m1', 'iranian-calendar-m2', 'iranian-calendar-m3',
 122                 'iranian-calendar-m4', 'iranian-calendar-m5', 'iranian-calendar-m6',
 123                 'iranian-calendar-m7', 'iranian-calendar-m8', 'iranian-calendar-m9',
 124                 'iranian-calendar-m10', 'iranian-calendar-m11', 'iranian-calendar-m12'
 125         ];
 126
 127         public static $mHebrewCalendarMonthMsgs = [
 128                 'hebrew-calendar-m1', 'hebrew-calendar-m2', 'hebrew-calendar-m3',
 129                 'hebrew-calendar-m4', 'hebrew-calendar-m5', 'hebrew-calendar-m6',
 130                 'hebrew-calendar-m7', 'hebrew-calendar-m8', 'hebrew-calendar-m9',
 131                 'hebrew-calendar-m10', 'hebrew-calendar-m11', 'hebrew-calendar-m12',
 132                 'hebrew-calendar-m6a', 'hebrew-calendar-m6b'
 133         ];
 134
 135         public static $mHebrewCalendarMonthGenMsgs = [
 136                 'hebrew-calendar-m1-gen', 'hebrew-calendar-m2-gen', 'hebrew-calendar-m3-gen',
 137                 'hebrew-calendar-m4-gen', 'hebrew-calendar-m5-gen', 'hebrew-calendar-m6-gen',
 138                 'hebrew-calendar-m7-gen', 'hebrew-calendar-m8-gen', 'hebrew-calendar-m9-gen',
 139                 'hebrew-calendar-m10-gen', 'hebrew-calendar-m11-gen', 'hebrew-calendar-m12-gen',
 140                 'hebrew-calendar-m6a-gen', 'hebrew-calendar-m6b-gen'
 141         ];
 142
 143         public static $mHijriCalendarMonthMsgs = [
 144                 'hijri-calendar-m1', 'hijri-calendar-m2', 'hijri-calendar-m3',
 145                 'hijri-calendar-m4', 'hijri-calendar-m5', 'hijri-calendar-m6',
 146                 'hijri-calendar-m7', 'hijri-calendar-m8', 'hijri-calendar-m9',
 147                 'hijri-calendar-m10', 'hijri-calendar-m11', 'hijri-calendar-m12'
 148         ];
 149
 150         /**
 151          * @since 1.20
 152          * @var array
 153          */
 154         public static $durationIntervals = [
 155                 'millennia' => 31556952000,
 156                 'centuries' => 3155695200,
 157                 'decades' => 315569520,
 158                 'years' => 31556952, // 86400 * ( 365 + ( 24 * 3 + 25 ) / 400 )
 159                 'weeks' => 604800,
 160                 'days' => 86400,
 161                 'hours' => 3600,
 162                 'minutes' => 60,
 163                 'seconds' => 1,
 164         ];
 165
 166         /**
 167          * Cache for language fallbacks.
 168          * @see Language::getFallbacksIncludingSiteLanguage
 169          * @since 1.21
 170          * @var array
 171          */
 172         private static $fallbackLanguageCache = [];
 173
 174         /**
 175          * Cache for grammar rules data
 176          * @var MapCacheLRU|null
 177          */
 178         private static $grammarTransformations;
 179
 180         /**
 181          * Cache for language names
 182          * @var HashBagOStuff|null
 183          */
 184         private static $languageNameCache;
 185
 186         /**
 187          * Unicode directional formatting characters, for embedBidi()
 188          */
 189         private static $lre = "\u{202A}"; // U+202A LEFT-TO-RIGHT EMBEDDING
 190         private static $rle = "\u{202B}"; // U+202B RIGHT-TO-LEFT EMBEDDING
 191         private static $pdf = "\u{202C}"; // U+202C POP DIRECTIONAL FORMATTING
 192
 193         /**
 194          * Directionality test regex for embedBidi(). Matches the first strong directionality codepoint:
 195          * - in group 1 if it is LTR
 196          * - in group 2 if it is RTL
 197          * Does not match if there is no strong directionality codepoint.
 198          *
 199          * The form is '/(?:([strong ltr codepoint])|([strong rtl codepoint]))/u' .
 200          *
 201          * Generated by UnicodeJS (see tools/strongDir) from the UCD; see
 202          * https://phabricator.wikimedia.org/diffusion/GUJS/ .
 203          */
 204         // @codeCoverageIgnoreStart
 205         // phpcs:ignore Generic.Files.LineLength
 206         private static $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';
 207         // @codeCoverageIgnoreEnd
 208
 209         /**
 210          * Get a cached or new language object for a given language code
 211          * @param string $code
 212          * @throws MWException
 213          * @return Language
 214          */
 215         static function factory( $code ) {
 216                 global $wgDummyLanguageCodes, $wgLangObjCacheSize;
 217
 218                 if ( isset( $wgDummyLanguageCodes[$code] ) ) {
 219                         $code = $wgDummyLanguageCodes[$code];
 220                 }
 221
 222                 // get the language object to process
 223                 $langObj = self::$mLangObjCache[$code] ?? self::newFromCode( $code );
 224
 225                 // merge the language object in to get it up front in the cache
 226                 self::$mLangObjCache = array_merge( [ $code => $langObj ], self::$mLangObjCache );
 227                 // get rid of the oldest ones in case we have an overflow
 228                 self::$mLangObjCache = array_slice( self::$mLangObjCache, 0, $wgLangObjCacheSize, true );
 229
 230                 return $langObj;
 231         }
 232
 233         /**
 234          * Create a language object for a given language code
 235          * @param string $code
 236          * @param bool $fallback Whether we're going through language fallback chain
 237          * @throws MWException
 238          * @return Language
 239          */
 240         protected static function newFromCode( $code, $fallback = false ) {
 241                 if ( !self::isValidCode( $code ) ) {
 242                         throw new MWException( "Invalid language code \"$code\"" );
 243                 }
 244
 245                 if ( !self::isValidBuiltInCode( $code ) ) {
 246                         // It's not possible to customise this code with class files, so
 247                         // just return a Language object. This is to support uselang= hacks.
 248                         $lang = new Language;
 249                         $lang->mCode = $code;
 250                         return $lang;
 251                 }
 252
 253                 // Check if there is a language class for the code
 254                 $class = self::classFromCode( $code, $fallback );
 255                 // LanguageCode does not inherit Language
 256                 if ( class_exists( $class ) && is_a( $class, 'Language', true ) ) {
 257                         $lang = new $class;
 258                         return $lang;
 259                 }
 260
 261                 // Keep trying the fallback list until we find an existing class
 262                 $fallbacks = self::getFallbacksFor( $code );
 263                 foreach ( $fallbacks as $fallbackCode ) {
 264                         if ( !self::isValidBuiltInCode( $fallbackCode ) ) {
 265                                 throw new MWException( "Invalid fallback '$fallbackCode' in fallback sequence for '$code'" );
 266                         }
 267
 268                         $class = self::classFromCode( $fallbackCode );
 269                         if ( class_exists( $class ) ) {
 270                                 $lang = new $class;
 271                                 $lang->mCode = $code;
 272                                 return $lang;
 273                         }
 274                 }
 275
 276                 throw new MWException( "Invalid fallback sequence for language '$code'" );
 277         }
 278
 279         /**
 280          * Intended for tests that may change configuration in a way that invalidates caches.
 281          *
 282          * @since 1.32
 283          */
 284         public static function clearCaches() {
 285                 if ( !defined( 'MW_PHPUNIT_TEST' ) ) {
 286                         throw new MWException( __METHOD__ . ' must not be used outside tests' );
 287                 }
 288                 self::$dataCache = null;
 289                 // Reinitialize $dataCache, since it's expected to always be available
 290                 self::getLocalisationCache();
 291                 self::$mLangObjCache = [];
 292                 self::$fallbackLanguageCache = [];
 293                 self::$grammarTransformations = null;
 294                 self::$languageNameCache = null;
 295         }
 296
 297         /**
 298          * Checks whether any localisation is available for that language tag
 299          * in MediaWiki (MessagesXx.php exists).
 300          *
 301          * @param string $code Language tag (in lower case)
 302          * @return bool Whether language is supported
 303          * @since 1.21
 304          */
 305         public static function isSupportedLanguage( $code ) {
 306                 if ( !self::isValidBuiltInCode( $code ) ) {
 307                         return false;
 308                 }
 309
 310                 if ( $code === 'qqq' ) {
 311                         return false;
 312                 }
 313
 314                 return is_readable( self::getMessagesFileName( $code ) ) ||
 315                         is_readable( self::getJsonMessagesFileName( $code ) );
 316         }
 317
 318         /**
 319          * Returns true if a language code string is a well-formed language tag
 320          * according to RFC 5646.
 321          * This function only checks well-formedness; it doesn't check that
 322          * language, script or variant codes actually exist in the repositories.
 323          *
 324          * Based on regexes by Mark Davis of the Unicode Consortium:
 325          * https://www.unicode.org/repos/cldr/trunk/tools/java/org/unicode/cldr/util/data/langtagRegex.txt
 326          *
 327          * @param string $code
 328          * @param bool $lenient Whether to allow '_' as separator. The default is only '-'.
 329          *
 330          * @return bool
 331          * @since 1.21
 332          */
 333         public static function isWellFormedLanguageTag( $code, $lenient = false ) {
 334                 $alpha = '[a-z]';
 335                 $digit = '[0-9]';
 336                 $alphanum = '[a-z0-9]';
 337                 $x = 'x'; # private use singleton
 338                 $singleton = '[a-wy-z]'; # other singleton
 339                 $s = $lenient ? '[-_]' : '-';
 340
 341                 $language = "$alpha{2,8}|$alpha{2,3}$s$alpha{3}";
 342                 $script = "$alpha{4}"; # ISO 15924
 343                 $region = "(?:$alpha{2}|$digit{3})"; # ISO 3166-1 alpha-2 or UN M.49
 344                 $variant = "(?:$alphanum{5,8}|$digit$alphanum{3})";
 345                 $extension = "$singleton(?:$s$alphanum{2,8})+";
 346                 $privateUse = "$x(?:$s$alphanum{1,8})+";
 347
 348                 # Define certain grandfathered codes, since otherwise the regex is pretty useless.
 349                 # Since these are limited, this is safe even later changes to the registry --
 350                 # the only oddity is that it might change the type of the tag, and thus
 351                 # the results from the capturing groups.
 352                 # https://www.iana.org/assignments/language-subtag-registry
 353
 354                 $grandfathered = "en{$s}GB{$s}oed"
 355                         . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
 356                         . "|no{$s}(?:bok|nyn)"
 357                         . "|sgn{$s}(?:BE{$s}(?:fr|nl)|CH{$s}de)"
 358                         . "|zh{$s}min{$s}nan";
 359
 360                 $variantList = "$variant(?:$s$variant)*";
 361                 $extensionList = "$extension(?:$s$extension)*";
 362
 363                 $langtag = "(?:($language)"
 364                         . "(?:$s$script)?"
 365                         . "(?:$s$region)?"
 366                         . "(?:$s$variantList)?"
 367                         . "(?:$s$extensionList)?"
 368                         . "(?:$s$privateUse)?)";
 369
 370                 # The final breakdown, with capturing groups for each of these components
 371                 # The variants, extensions, grandfathered, and private-use may have interior '-'
 372
 373                 $root = "^(?:$langtag|$privateUse|$grandfathered)$";
 374
 375                 return (bool)preg_match( "/$root/", strtolower( $code ) );
 376         }
 377
 378         /**
 379          * Returns true if a language code string is of a valid form, whether or
 380          * not it exists. This includes codes which are used solely for
 381          * customisation via the MediaWiki namespace.
 382          *
 383          * @param string $code
 384          *
 385          * @return bool
 386          */
 387         public static function isValidCode( $code ) {
 388                 static $cache = [];
 389                 if ( !isset( $cache[$code] ) ) {
 390                         // People think language codes are html safe, so enforce it.
 391                         // Ideally we should only allow a-zA-Z0-9-
 392                         // but, .+ and other chars are often used for {{int:}} hacks
 393                         // see bugs T39564, T39587, T38938
 394                         $cache[$code] =
 395                                 // Protect against path traversal
 396                                 strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
 397                                 && !preg_match( MediaWikiTitleCodec::getTitleInvalidRegex(), $code );
 398                 }
 399                 return $cache[$code];
 400         }
 401
 402         /**
 403          * Returns true if a language code is of a valid form for the purposes of
 404          * internal customisation of MediaWiki, via Messages*.php or *.json.
 405          *
 406          * @param string $code
 407          *
 408          * @since 1.18
 409          * @return bool
 410          */
 411         public static function isValidBuiltInCode( $code ) {
 412                 Assert::parameterType( 'string', $code, '$code' );
 413
 414                 return (bool)preg_match( '/^[a-z0-9-]{2,}$/', $code );
 415         }
 416
 417         /**
 418          * Returns true if a language code is an IETF tag known to MediaWiki.
 419          *
 420          * @param string $tag
 421          *
 422          * @since 1.21
 423          * @return bool
 424          */
 425         public static function isKnownLanguageTag( $tag ) {
 426                 // Quick escape for invalid input to avoid exceptions down the line
 427                 // when code tries to process tags which are not valid at all.
 428                 if ( !self::isValidBuiltInCode( $tag ) ) {
 429                         return false;
 430                 }
 431
 432                 if ( isset( MediaWiki\Languages\Data\Names::$names[$tag] )
 433                         || self::fetchLanguageName( $tag, $tag ) !== ''
 434                 ) {
 435                         return true;
 436                 }
 437
 438                 return false;
 439         }
 440
 441         /**
 442          * Get the LocalisationCache instance
 443          *
 444          * @return LocalisationCache
 445          */
 446         public static function getLocalisationCache() {
 447                 if ( is_null( self::$dataCache ) ) {
 448                         global $wgLocalisationCacheConf;
 449                         $class = $wgLocalisationCacheConf['class'];
 450                         self::$dataCache = new $class( $wgLocalisationCacheConf );
 451                 }
 452                 return self::$dataCache;
 453         }
 454
 455         function __construct() {
 456                 $this->mConverter = new FakeConverter( $this );
 457                 // Set the code to the name of the descendant
 458                 if ( static::class === 'Language' ) {
 459                         $this->mCode = 'en';
 460                 } else {
 461                         $this->mCode = str_replace( '_', '-', strtolower( substr( static::class, 8 ) ) );
 462                 }
 463                 self::getLocalisationCache();
 464         }
 465
 466         /**
 467          * Reduce memory usage
 468          */
 469         function __destruct() {
 470                 foreach ( $this as $name => $value ) {
 471                         unset( $this->$name );
 472                 }
 473         }
 474
 475         /**
 476          * Hook which will be called if this is the content language.
 477          * Descendants can use this to register hook functions or modify globals
 478          */
 479         function initContLang() {
 480         }
 481
 482         /**
 483          * @return array
 484          * @since 1.19
 485          */
 486         public function getFallbackLanguages() {
 487                 return self::getFallbacksFor( $this->mCode );
 488         }
 489
 490         /**
 491          * Exports $wgBookstoreListEn
 492          * @return array
 493          */
 494         public function getBookstoreList() {
 495                 return self::$dataCache->getItem( $this->mCode, 'bookstoreList' );
 496         }
 497
 498         /**
 499          * Returns an array of localised namespaces indexed by their numbers. If the namespace is not
 500          * available in localised form, it will be included in English.
 501          *
 502          * @return string[] List of localized namespace names, indexed by numeric namespace ID.
 503          */
 504         public function getNamespaces() {
 505                 if ( is_null( $this->namespaceNames ) ) {
 506                         global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 507
 508                         $validNamespaces = MWNamespace::getCanonicalNamespaces();
 509
 510                         $this->namespaceNames = $wgExtraNamespaces +
 511                                 self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
 512                         $this->namespaceNames += $validNamespaces;
 513
 514                         $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
 515                         if ( $wgMetaNamespaceTalk ) {
 516                                 $this->namespaceNames[NS_PROJECT_TALK] = $wgMetaNamespaceTalk;
 517                         } else {
 518                                 $talk = $this->namespaceNames[NS_PROJECT_TALK];
 519                                 $this->namespaceNames[NS_PROJECT_TALK] =
 520                                         $this->fixVariableInNamespace( $talk );
 521                         }
 522
 523                         # Sometimes a language will be localised but not actually exist on this wiki.
 524                         foreach ( $this->namespaceNames as $key => $text ) {
 525                                 if ( !isset( $validNamespaces[$key] ) ) {
 526                                         unset( $this->namespaceNames[$key] );
 527                                 }
 528                         }
 529
 530                         # The above mixing may leave namespaces out of canonical order.
 531                         # Re-order by namespace ID number...
 532                         ksort( $this->namespaceNames );
 533
 534                         Hooks::run( 'LanguageGetNamespaces', [ &$this->namespaceNames ] );
 535                 }
 536
 537                 return $this->namespaceNames;
 538         }
 539
 540         /**
 541          * Arbitrarily set all of the namespace names at once. Mainly used for testing
 542          * @param array $namespaces Array of namespaces (id => name)
 543          */
 544         public function setNamespaces( array $namespaces ) {
 545                 $this->namespaceNames = $namespaces;
 546                 $this->mNamespaceIds = null;
 547         }
 548
 549         /**
 550          * Resets all of the namespace caches. Mainly used for testing
 551          */
 552         public function resetNamespaces() {
 553                 $this->namespaceNames = null;
 554                 $this->mNamespaceIds = null;
 555                 $this->namespaceAliases = null;
 556         }
 557
 558         /**
 559          * A convenience function that returns getNamespaces() with spaces instead of underscores
 560          * in values. Useful for producing output to be displayed e.g. in `<select>` forms.
 561          *
 562          * @return array
 563          */
 564         public function getFormattedNamespaces() {
 565                 $ns = $this->getNamespaces();
 566                 foreach ( $ns as $k => $v ) {
 567                         $ns[$k] = strtr( $v, '_', ' ' );
 568                 }
 569                 return $ns;
 570         }
 571
 572         /**
 573          * Get a namespace value by key
 574          *
 575          * <code>
 576          * $mw_ns = $lang->getNsText( NS_MEDIAWIKI );
 577          * echo $mw_ns; // prints 'MediaWiki'
 578          * </code>
 579          *
 580          * @param int $index The array key of the namespace to return
 581          * @return string|bool String if the namespace value exists, otherwise false
 582          */
 583         public function getNsText( $index ) {
 584                 $ns = $this->getNamespaces();
 585                 return $ns[$index] ?? false;
 586         }
 587
 588         /**
 589          * A convenience function that returns the same thing as
 590          * getNsText() except with '_' changed to ' ', useful for
 591          * producing output.
 592          *
 593          * <code>
 594          * $mw_ns = $lang->getFormattedNsText( NS_MEDIAWIKI_TALK );
 595          * echo $mw_ns; // prints 'MediaWiki talk'
 596          * </code>
 597          *
 598          * @param int $index The array key of the namespace to return
 599          * @return string Namespace name without underscores (empty string if namespace does not exist)
 600          */
 601         public function getFormattedNsText( $index ) {
 602                 $ns = $this->getNsText( $index );
 603                 return strtr( $ns, '_', ' ' );
 604         }
 605
 606         /**
 607          * Returns gender-dependent namespace alias if available.
 608          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 609          * @param int $index Namespace index
 610          * @param string $gender Gender key (male, female... )
 611          * @return string
 612          * @since 1.18
 613          */
 614         public function getGenderNsText( $index, $gender ) {
 615                 global $wgExtraGenderNamespaces;
 616
 617                 $ns = $wgExtraGenderNamespaces +
 618                         (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 619
 620                 return $ns[$index][$gender] ?? $this->getNsText( $index );
 621         }
 622
 623         /**
 624          * Whether this language uses gender-dependent namespace aliases.
 625          * See https://www.mediawiki.org/wiki/Manual:$wgExtraGenderNamespaces
 626          * @return bool
 627          * @since 1.18
 628          */
 629         public function needsGenderDistinction() {
 630                 global $wgExtraGenderNamespaces, $wgExtraNamespaces;
 631                 if ( count( $wgExtraGenderNamespaces ) > 0 ) {
 632                         // $wgExtraGenderNamespaces overrides everything
 633                         return true;
 634                 } elseif ( isset( $wgExtraNamespaces[NS_USER] ) && isset( $wgExtraNamespaces[NS_USER_TALK] ) ) {
 635                         /// @todo There may be other gender namespace than NS_USER & NS_USER_TALK in the future
 636                         // $wgExtraNamespaces overrides any gender aliases specified in i18n files
 637                         return false;
 638                 } else {
 639                         // Check what is in i18n files
 640                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 641                         return count( $aliases ) > 0;
 642                 }
 643         }
 644
 645         /**
 646          * Get a namespace key by value, case insensitive.
 647          * Only matches namespace names for the current language, not the
 648          * canonical ones defined in Namespace.php.
 649          *
 650          * @param string $text
 651          * @return int|bool An integer if $text is a valid value otherwise false
 652          */
 653         function getLocalNsIndex( $text ) {
 654                 $lctext = $this->lc( $text );
 655                 $ids = $this->getNamespaceIds();
 656                 return $ids[$lctext] ?? false;
 657         }
 658
 659         /**
 660          * @return array
 661          */
 662         public function getNamespaceAliases() {
 663                 if ( is_null( $this->namespaceAliases ) ) {
 664                         $aliases = self::$dataCache->getItem( $this->mCode, 'namespaceAliases' );
 665                         if ( !$aliases ) {
 666                                 $aliases = [];
 667                         } else {
 668                                 foreach ( $aliases as $name => $index ) {
 669                                         if ( $index === NS_PROJECT_TALK ) {
 670                                                 unset( $aliases[$name] );
 671                                                 $name = $this->fixVariableInNamespace( $name );
 672                                                 $aliases[$name] = $index;
 673                                         }
 674                                 }
 675                         }
 676
 677                         global $wgExtraGenderNamespaces;
 678                         $genders = $wgExtraGenderNamespaces +
 679                                 (array)self::$dataCache->getItem( $this->mCode, 'namespaceGenderAliases' );
 680                         foreach ( $genders as $index => $forms ) {
 681                                 foreach ( $forms as $alias ) {
 682                                         $aliases[$alias] = $index;
 683                                 }
 684                         }
 685
 686                         # Also add converted namespace names as aliases, to avoid confusion.
 687                         $convertedNames = [];
 688                         foreach ( $this->getVariants() as $variant ) {
 689                                 if ( $variant === $this->mCode ) {
 690                                         continue;
 691                                 }
 692                                 foreach ( $this->getNamespaces() as $ns => $_ ) {
 693                                         $convertedNames[$this->getConverter()->convertNamespace( $ns, $variant )] = $ns;
 694                                 }
 695                         }
 696
 697                         $this->namespaceAliases = $aliases + $convertedNames;
 698
 699                         # Filter out aliases to namespaces that don't exist, e.g. from extensions
 700                         # that aren't loaded here but are included in the l10n cache.
 701                         # (array_intersect preserves keys from its first argument)
 702                         $this->namespaceAliases = array_intersect(
 703                                 $this->namespaceAliases,
 704                                 array_keys( $this->getNamespaces() )
 705                         );
 706                 }
 707
 708                 return $this->namespaceAliases;
 709         }
 710
 711         /**
 712          * @return array
 713          */
 714         public function getNamespaceIds() {
 715                 if ( is_null( $this->mNamespaceIds ) ) {
 716                         global $wgNamespaceAliases;
 717                         # Put namespace names and aliases into a hashtable.
 718                         # If this is too slow, then we should arrange it so that it is done
 719                         # before caching. The catch is that at pre-cache time, the above
 720                         # class-specific fixup hasn't been done.
 721                         $this->mNamespaceIds = [];
 722                         foreach ( $this->getNamespaces() as $index => $name ) {
 723                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 724                         }
 725                         foreach ( $this->getNamespaceAliases() as $name => $index ) {
 726                                 $this->mNamespaceIds[$this->lc( $name )] = $index;
 727                         }
 728                         if ( $wgNamespaceAliases ) {
 729                                 foreach ( $wgNamespaceAliases as $name => $index ) {
 730                                         $this->mNamespaceIds[$this->lc( $name )] = $index;
 731                                 }
 732                         }
 733                 }
 734                 return $this->mNamespaceIds;
 735         }
 736
 737         /**
 738          * Get a namespace key by value, case insensitive.  Canonical namespace
 739          * names override custom ones defined for the current language.
 740          *
 741          * @param string $text
 742          * @return int|bool An integer if $text is a valid value otherwise false
 743          */
 744         public function getNsIndex( $text ) {
 745                 $lctext = $this->lc( $text );
 746                 $ns = MWNamespace::getCanonicalIndex( $lctext );
 747                 if ( $ns !== null ) {
 748                         return $ns;
 749                 }
 750                 $ids = $this->getNamespaceIds();
 751                 return $ids[$lctext] ?? false;
 752         }
 753
 754         /**
 755          * short names for language variants used for language conversion links.
 756          *
 757          * @param string $code
 758          * @param bool $usemsg Use the "variantname-xyz" message if it exists
 759          * @return string
 760          */
 761         public function getVariantname( $code, $usemsg = true ) {
 762                 $msg = "variantname-$code";
 763                 if ( $usemsg && wfMessage( $msg )->exists() ) {
 764                         return $this->getMessageFromDB( $msg );
 765                 }
 766                 $name = self::fetchLanguageName( $code );
 767                 if ( $name ) {
 768                         return $name; # if it's defined as a language name, show that
 769                 } else {
 770                         # otherwise, output the language code
 771                         return $code;
 772                 }
 773         }
 774
 775         /**
 776          * @return string[]|bool List of date format preference keys, or false if disabled.
 777          */
 778         public function getDatePreferences() {
 779                 return self::$dataCache->getItem( $this->mCode, 'datePreferences' );
 780         }
 781
 782         /**
 783          * @return array
 784          */
 785         function getDateFormats() {
 786                 return self::$dataCache->getItem( $this->mCode, 'dateFormats' );
 787         }
 788
 789         /**
 790          * @return array|string
 791          */
 792         public function getDefaultDateFormat() {
 793                 $df = self::$dataCache->getItem( $this->mCode, 'defaultDateFormat' );
 794                 if ( $df === 'dmy or mdy' ) {
 795                         global $wgAmericanDates;
 796                         return $wgAmericanDates ? 'mdy' : 'dmy';
 797                 } else {
 798                         return $df;
 799                 }
 800         }
 801
 802         /**
 803          * @return array
 804          */
 805         public function getDatePreferenceMigrationMap() {
 806                 return self::$dataCache->getItem( $this->mCode, 'datePreferenceMigrationMap' );
 807         }
 808
 809         /**
 810          * @return array
 811          */
 812         public function getExtraUserToggles() {
 813                 return (array)self::$dataCache->getItem( $this->mCode, 'extraUserToggles' );
 814         }
 815
 816         /**
 817          * @param string $tog
 818          * @return string
 819          */
 820         function getUserToggle( $tog ) {
 821                 return $this->getMessageFromDB( "tog-$tog" );
 822         }
 823
 824         /**
 825          * Get an array of language names, indexed by code.
 826          * @param null|string $inLanguage Code of language in which to return the names
 827          *              Use self::AS_AUTONYMS for autonyms (native names)
 828          * @param string $include One of:
 829          *              self::ALL all available languages
 830          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 831          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 832          * @return array Language code => language name (sorted by key)
 833          * @since 1.20
 834          */
 835         public static function fetchLanguageNames( $inLanguage = self::AS_AUTONYMS, $include = 'mw' ) {
 836                 $cacheKey = $inLanguage === self::AS_AUTONYMS ? 'null' : $inLanguage;
 837                 $cacheKey .= ":$include";
 838                 if ( self::$languageNameCache === null ) {
 839                         self::$languageNameCache = new HashBagOStuff( [ 'maxKeys' => 20 ] );
 840                 }
 841
 842                 $ret = self::$languageNameCache->get( $cacheKey );
 843                 if ( !$ret ) {
 844                         $ret = self::fetchLanguageNamesUncached( $inLanguage, $include );
 845                         self::$languageNameCache->set( $cacheKey, $ret );
 846                 }
 847                 return $ret;
 848         }
 849
 850         /**
 851          * Uncached helper for fetchLanguageNames
 852          * @param null|string $inLanguage Code of language in which to return the names
 853          *              Use self::AS_AUTONYMS for autonyms (native names)
 854          * @param string $include One of:
 855          *              self::ALL all available languages
 856          *              'mw' only if the language is defined in MediaWiki or wgExtraLanguageNames (default)
 857          *              self::SUPPORTED only if the language is in 'mw' *and* has a message file
 858          * @return array Language code => language name (sorted by key)
 859          */
 860         private static function fetchLanguageNamesUncached(
 861                 $inLanguage = self::AS_AUTONYMS,
 862                 $include = 'mw'
 863         ) {
 864                 global $wgExtraLanguageNames, $wgUsePigLatinVariant;
 865
 866                 // If passed an invalid language code to use, fallback to en
 867                 if ( $inLanguage !== self::AS_AUTONYMS && !self::isValidCode( $inLanguage ) ) {
 868                         $inLanguage = 'en';
 869                 }
 870
 871                 $names = [];
 872
 873                 if ( $inLanguage ) {
 874                         # TODO: also include when $inLanguage is null, when this code is more efficient
 875                         Hooks::run( 'LanguageGetTranslatedLanguageNames', [ &$names, $inLanguage ] );
 876                 }
 877
 878                 $mwNames = $wgExtraLanguageNames + MediaWiki\Languages\Data\Names::$names;
 879                 if ( $wgUsePigLatinVariant ) {
 880                         // Pig Latin (for variant development)
 881                         $mwNames['en-x-piglatin'] = 'Igpay Atinlay';
 882                 }
 883
 884                 foreach ( $mwNames as $mwCode => $mwName ) {
 885                         # - Prefer own MediaWiki native name when not using the hook
 886                         # - For other names just add if not added through the hook
 887                         if ( $mwCode === $inLanguage || !isset( $names[$mwCode] ) ) {
 888                                 $names[$mwCode] = $mwName;
 889                         }
 890                 }
 891
 892                 if ( $include === self::ALL ) {
 893                         ksort( $names );
 894                         return $names;
 895                 }
 896
 897                 $returnMw = [];
 898                 $coreCodes = array_keys( $mwNames );
 899                 foreach ( $coreCodes as $coreCode ) {
 900                         $returnMw[$coreCode] = $names[$coreCode];
 901                 }
 902
 903                 if ( $include === self::SUPPORTED ) {
 904                         $namesMwFile = [];
 905                         # We do this using a foreach over the codes instead of a directory
 906                         # loop so that messages files in extensions will work correctly.
 907                         foreach ( $returnMw as $code => $value ) {
 908                                 if ( is_readable( self::getMessagesFileName( $code ) )
 909                                         || is_readable( self::getJsonMessagesFileName( $code ) )
 910                                 ) {
 911                                         $namesMwFile[$code] = $names[$code];
 912                                 }
 913                         }
 914
 915                         ksort( $namesMwFile );
 916                         return $namesMwFile;
 917                 }
 918
 919                 ksort( $returnMw );
 920                 # 'mw' option; default if it's not one of the other two options (all/mwfile)
 921                 return $returnMw;
 922         }
 923
 924         /**
 925          * @param string $code The code of the language for which to get the name
 926          * @param null|string $inLanguage Code of language in which to return the name
 927          *   (SELF::AS_AUTONYMS for autonyms)
 928          * @param string $include See fetchLanguageNames()
 929          * @return string Language name or empty
 930          * @since 1.20
 931          */
 932         public static function fetchLanguageName(
 933                 $code,
 934                 $inLanguage = self::AS_AUTONYMS,
 935                 $include = self::ALL
 936         ) {
 937                 $code = strtolower( $code );
 938                 $array = self::fetchLanguageNames( $inLanguage, $include );
 939                 return !array_key_exists( $code, $array ) ? '' : $array[$code];
 940         }
 941
 942         /**
 943          * Get a message from the MediaWiki namespace.
 944          *
 945          * @param string $msg Message name
 946          * @return string
 947          */
 948         public function getMessageFromDB( $msg ) {
 949                 return $this->msg( $msg )->text();
 950         }
 951
 952         /**
 953          * Get message object in this language. Only for use inside this class.
 954          *
 955          * @param string $msg Message name
 956          * @return Message
 957          */
 958         protected function msg( $msg ) {
 959                 return wfMessage( $msg )->inLanguage( $this );
 960         }
 961
 962         /**
 963          * @param string $key
 964          * @return string
 965          */
 966         public function getMonthName( $key ) {
 967                 return $this->getMessageFromDB( self::$mMonthMsgs[$key - 1] );
 968         }
 969
 970         /**
 971          * @return array
 972          */
 973         public function getMonthNamesArray() {
 974                 $monthNames = [ '' ];
 975                 for ( $i = 1; $i < 13; $i++ ) {
 976                         $monthNames[] = $this->getMonthName( $i );
 977                 }
 978                 return $monthNames;
 979         }
 980
 981         /**
 982          * @param string $key
 983          * @return string
 984          */
 985         public function getMonthNameGen( $key ) {
 986                 return $this->getMessageFromDB( self::$mMonthGenMsgs[$key - 1] );
 987         }
 988
 989         /**
 990          * @param string $key
 991          * @return string
 992          */
 993         public function getMonthAbbreviation( $key ) {
 994                 return $this->getMessageFromDB( self::$mMonthAbbrevMsgs[$key - 1] );
 995         }
 996
 997         /**
 998          * @return array
 999          */
1000         public function getMonthAbbreviationsArray() {
1001                 $monthNames = [ '' ];
1002                 for ( $i = 1; $i < 13; $i++ ) {
1003                         $monthNames[] = $this->getMonthAbbreviation( $i );
1004                 }
1005                 return $monthNames;
1006         }
1007
1008         /**
1009          * @param string $key
1010          * @return string
1011          */
1012         public function getWeekdayName( $key ) {
1013                 return $this->getMessageFromDB( self::$mWeekdayMsgs[$key - 1] );
1014         }
1015
1016         /**
1017          * @param string $key
1018          * @return string
1019          */
1020         function getWeekdayAbbreviation( $key ) {
1021                 return $this->getMessageFromDB( self::$mWeekdayAbbrevMsgs[$key - 1] );
1022         }
1023
1024         /**
1025          * @param string $key
1026          * @return string
1027          */
1028         function getIranianCalendarMonthName( $key ) {
1029                 return $this->getMessageFromDB( self::$mIranianCalendarMonthMsgs[$key - 1] );
1030         }
1031
1032         /**
1033          * @param string $key
1034          * @return string
1035          */
1036         function getHebrewCalendarMonthName( $key ) {
1037                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthMsgs[$key - 1] );
1038         }
1039
1040         /**
1041          * @param string $key
1042          * @return string
1043          */
1044         function getHebrewCalendarMonthNameGen( $key ) {
1045                 return $this->getMessageFromDB( self::$mHebrewCalendarMonthGenMsgs[$key - 1] );
1046         }
1047
1048         /**
1049          * @param string $key
1050          * @return string
1051          */
1052         function getHijriCalendarMonthName( $key ) {
1053                 return $this->getMessageFromDB( self::$mHijriCalendarMonthMsgs[$key - 1] );
1054         }
1055
1056         /**
1057          * Pass through result from $dateTimeObj->format()
1058          * @param DateTime|bool|null &$dateTimeObj
1059          * @param string $ts
1060          * @param DateTimeZone|bool|null $zone
1061          * @param string $code
1062          * @return string
1063          */
1064         private static function dateTimeObjFormat( &$dateTimeObj, $ts, $zone, $code ) {
1065                 if ( !$dateTimeObj ) {
1066                         $dateTimeObj = DateTime::createFromFormat(
1067                                 'YmdHis', $ts, $zone ?: new DateTimeZone( 'UTC' )
1068                         );
1069                 }
1070                 return $dateTimeObj->format( $code );
1071         }
1072
1073         /**
1074          * This is a workalike of PHP's date() function, but with better
1075          * internationalisation, a reduced set of format characters, and a better
1076          * escaping format.
1077          *
1078          * Supported format characters are dDjlNwzWFmMntLoYyaAgGhHiscrUeIOPTZ. See
1079          * the PHP manual for definitions. There are a number of extensions, which
1080          * start with "x":
1081          *
1082          *    xn   Do not translate digits of the next numeric format character
1083          *    xN   Toggle raw digit (xn) flag, stays set until explicitly unset
1084          *    xr   Use roman numerals for the next numeric format character
1085          *    xh   Use hebrew numerals for the next numeric format character
1086          *    xx   Literal x
1087          *    xg   Genitive month name
1088          *
1089          *    xij  j (day number) in Iranian calendar
1090          *    xiF  F (month name) in Iranian calendar
1091          *    xin  n (month number) in Iranian calendar
1092          *    xiy  y (two digit year) in Iranian calendar
1093          *    xiY  Y (full year) in Iranian calendar
1094          *    xit  t (days in month) in Iranian calendar
1095          *    xiz  z (day of the year) in Iranian calendar
1096          *
1097          *    xjj  j (day number) in Hebrew calendar
1098          *    xjF  F (month name) in Hebrew calendar
1099          *    xjt  t (days in month) in Hebrew calendar
1100          *    xjx  xg (genitive month name) in Hebrew calendar
1101          *    xjn  n (month number) in Hebrew calendar
1102          *    xjY  Y (full year) in Hebrew calendar
1103          *
1104          *    xmj  j (day number) in Hijri calendar
1105          *    xmF  F (month name) in Hijri calendar
1106          *    xmn  n (month number) in Hijri calendar
1107          *    xmY  Y (full year) in Hijri calendar
1108          *
1109          *    xkY  Y (full year) in Thai solar calendar. Months and days are
1110          *                       identical to the Gregorian calendar
1111          *    xoY  Y (full year) in Minguo calendar or Juche year.
1112          *                       Months and days are identical to the
1113          *                       Gregorian calendar
1114          *    xtY  Y (full year) in Japanese nengo. Months and days are
1115          *                       identical to the Gregorian calendar
1116          *
1117          * Characters enclosed in double quotes will be considered literal (with
1118          * the quotes themselves removed). Unmatched quotes will be considered
1119          * literal quotes. Example:
1120          *
1121          * "The month is" F       => The month is January
1122          * i's"                   => 20'11"
1123          *
1124          * Backslash escaping is also supported.
1125          *
1126          * Input timestamp is assumed to be pre-normalized to the desired local
1127          * time zone, if any. Note that the format characters crUeIOPTZ will assume
1128          * $ts is UTC if $zone is not given.
1129          *
1130          * @param string $format
1131          * @param string $ts 14-character timestamp
1132          *      YYYYMMDDHHMMSS
1133          *      01234567890123
1134          * @param DateTimeZone|null $zone Timezone of $ts
1135          * @param int &$ttl The amount of time (in seconds) the output may be cached for.
1136          * Only makes sense if $ts is the current time.
1137          * @todo handling of "o" format character for Iranian, Hebrew, Hijri & Thai?
1138          *
1139          * @throws MWException
1140          * @return string
1141          */
1142         public function sprintfDate( $format, $ts, DateTimeZone $zone = null, &$ttl = 'unused' ) {
1143                 $s = '';
1144                 $raw = false;
1145                 $roman = false;
1146                 $hebrewNum = false;
1147                 $dateTimeObj = false;
1148                 $rawToggle = false;
1149                 $iranian = false;
1150                 $hebrew = false;
1151                 $hijri = false;
1152                 $thai = false;
1153                 $minguo = false;
1154                 $tenno = false;
1155
1156                 $usedSecond = false;
1157                 $usedMinute = false;
1158                 $usedHour = false;
1159                 $usedAMPM = false;
1160                 $usedDay = false;
1161                 $usedWeek = false;
1162                 $usedMonth = false;
1163                 $usedYear = false;
1164                 $usedISOYear = false;
1165                 $usedIsLeapYear = false;
1166
1167                 $usedHebrewMonth = false;
1168                 $usedIranianMonth = false;
1169                 $usedHijriMonth = false;
1170                 $usedHebrewYear = false;
1171                 $usedIranianYear = false;
1172                 $usedHijriYear = false;
1173                 $usedTennoYear = false;
1174
1175                 if ( strlen( $ts ) !== 14 ) {
1176                         throw new MWException( __METHOD__ . ": The timestamp $ts should have 14 characters" );
1177                 }
1178
1179                 if ( !ctype_digit( $ts ) ) {
1180                         throw new MWException( __METHOD__ . ": The timestamp $ts should be a number" );
1181                 }
1182
1183                 $formatLength = strlen( $format );
1184                 for ( $p = 0; $p < $formatLength; $p++ ) {
1185                         $num = false;
1186                         $code = $format[$p];
1187                         if ( $code == 'x' && $p < $formatLength - 1 ) {
1188                                 $code .= $format[++$p];
1189                         }
1190
1191                         if ( ( $code === 'xi'
1192                                         || $code === 'xj'
1193                                         || $code === 'xk'
1194                                         || $code === 'xm'
1195                                         || $code === 'xo'
1196                                         || $code === 'xt' )
1197                                 && $p < $formatLength - 1 ) {
1198                                 $code .= $format[++$p];
1199                         }
1200
1201                         switch ( $code ) {
1202                                 case 'xx':
1203                                         $s .= 'x';
1204                                         break;
1205                                 case 'xn':
1206                                         $raw = true;
1207                                         break;
1208                                 case 'xN':
1209                                         $rawToggle = !$rawToggle;
1210                                         break;
1211                                 case 'xr':
1212                                         $roman = true;
1213                                         break;
1214                                 case 'xh':
1215                                         $hebrewNum = true;
1216                                         break;
1217                                 case 'xg':
1218                                         $usedMonth = true;
1219                                         $s .= $this->getMonthNameGen( substr( $ts, 4, 2 ) );
1220                                         break;
1221                                 case 'xjx':
1222                                         $usedHebrewMonth = true;
1223                                         if ( !$hebrew ) {
1224                                                 $hebrew = self::tsToHebrew( $ts );
1225                                         }
1226                                         $s .= $this->getHebrewCalendarMonthNameGen( $hebrew[1] );
1227                                         break;
1228                                 case 'd':
1229                                         $usedDay = true;
1230                                         $num = substr( $ts, 6, 2 );
1231                                         break;
1232                                 case 'D':
1233                                         $usedDay = true;
1234                                         $s .= $this->getWeekdayAbbreviation(
1235                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1236                                         );
1237                                         break;
1238                                 case 'j':
1239                                         $usedDay = true;
1240                                         $num = intval( substr( $ts, 6, 2 ) );
1241                                         break;
1242                                 case 'xij':
1243                                         $usedDay = true;
1244                                         if ( !$iranian ) {
1245                                                 $iranian = self::tsToIranian( $ts );
1246                                         }
1247                                         $num = $iranian[2];
1248                                         break;
1249                                 case 'xmj':
1250                                         $usedDay = true;
1251                                         if ( !$hijri ) {
1252                                                 $hijri = self::tsToHijri( $ts );
1253                                         }
1254                                         $num = $hijri[2];
1255                                         break;
1256                                 case 'xjj':
1257                                         $usedDay = true;
1258                                         if ( !$hebrew ) {
1259                                                 $hebrew = self::tsToHebrew( $ts );
1260                                         }
1261                                         $num = $hebrew[2];
1262                                         break;
1263                                 case 'l':
1264                                         $usedDay = true;
1265                                         $s .= $this->getWeekdayName(
1266                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'w' ) + 1
1267                                         );
1268                                         break;
1269                                 case 'F':
1270                                         $usedMonth = true;
1271                                         $s .= $this->getMonthName( substr( $ts, 4, 2 ) );
1272                                         break;
1273                                 case 'xiF':
1274                                         $usedIranianMonth = true;
1275                                         if ( !$iranian ) {
1276                                                 $iranian = self::tsToIranian( $ts );
1277                                         }
1278                                         $s .= $this->getIranianCalendarMonthName( $iranian[1] );
1279                                         break;
1280                                 case 'xmF':
1281                                         $usedHijriMonth = true;
1282                                         if ( !$hijri ) {
1283                                                 $hijri = self::tsToHijri( $ts );
1284                                         }
1285                                         $s .= $this->getHijriCalendarMonthName( $hijri[1] );
1286                                         break;
1287                                 case 'xjF':
1288                                         $usedHebrewMonth = true;
1289                                         if ( !$hebrew ) {
1290                                                 $hebrew = self::tsToHebrew( $ts );
1291                                         }
1292                                         $s .= $this->getHebrewCalendarMonthName( $hebrew[1] );
1293                                         break;
1294                                 case 'm':
1295                                         $usedMonth = true;
1296                                         $num = substr( $ts, 4, 2 );
1297                                         break;
1298                                 case 'M':
1299                                         $usedMonth = true;
1300                                         $s .= $this->getMonthAbbreviation( substr( $ts, 4, 2 ) );
1301                                         break;
1302                                 case 'n':
1303                                         $usedMonth = true;
1304                                         $num = intval( substr( $ts, 4, 2 ) );
1305                                         break;
1306                                 case 'xin':
1307                                         $usedIranianMonth = true;
1308                                         if ( !$iranian ) {
1309                                                 $iranian = self::tsToIranian( $ts );
1310                                         }
1311                                         $num = $iranian[1];
1312                                         break;
1313                                 case 'xmn':
1314                                         $usedHijriMonth = true;
1315                                         if ( !$hijri ) {
1316                                                 $hijri = self::tsToHijri( $ts );
1317                                         }
1318                                         $num = $hijri[1];
1319                                         break;
1320                                 case 'xjn':
1321                                         $usedHebrewMonth = true;
1322                                         if ( !$hebrew ) {
1323                                                 $hebrew = self::tsToHebrew( $ts );
1324                                         }
1325                                         $num = $hebrew[1];
1326                                         break;
1327                                 case 'xjt':
1328                                         $usedHebrewMonth = true;
1329                                         if ( !$hebrew ) {
1330                                                 $hebrew = self::tsToHebrew( $ts );
1331                                         }
1332                                         $num = $hebrew[3];
1333                                         break;
1334                                 case 'Y':
1335                                         $usedYear = true;
1336                                         $num = substr( $ts, 0, 4 );
1337                                         break;
1338                                 case 'xiY':
1339                                         $usedIranianYear = true;
1340                                         if ( !$iranian ) {
1341                                                 $iranian = self::tsToIranian( $ts );
1342                                         }
1343                                         $num = $iranian[0];
1344                                         break;
1345                                 case 'xmY':
1346                                         $usedHijriYear = true;
1347                                         if ( !$hijri ) {
1348                                                 $hijri = self::tsToHijri( $ts );
1349                                         }
1350                                         $num = $hijri[0];
1351                                         break;
1352                                 case 'xjY':
1353                                         $usedHebrewYear = true;
1354                                         if ( !$hebrew ) {
1355                                                 $hebrew = self::tsToHebrew( $ts );
1356                                         }
1357                                         $num = $hebrew[0];
1358                                         break;
1359                                 case 'xkY':
1360                                         $usedYear = true;
1361                                         if ( !$thai ) {
1362                                                 $thai = self::tsToYear( $ts, 'thai' );
1363                                         }
1364                                         $num = $thai[0];
1365                                         break;
1366                                 case 'xoY':
1367                                         $usedYear = true;
1368                                         if ( !$minguo ) {
1369                                                 $minguo = self::tsToYear( $ts, 'minguo' );
1370                                         }
1371                                         $num = $minguo[0];
1372                                         break;
1373                                 case 'xtY':
1374                                         $usedTennoYear = true;
1375                                         if ( !$tenno ) {
1376                                                 $tenno = self::tsToYear( $ts, 'tenno' );
1377                                         }
1378                                         $num = $tenno[0];
1379                                         break;
1380                                 case 'y':
1381                                         $usedYear = true;
1382                                         $num = substr( $ts, 2, 2 );
1383                                         break;
1384                                 case 'xiy':
1385                                         $usedIranianYear = true;
1386                                         if ( !$iranian ) {
1387                                                 $iranian = self::tsToIranian( $ts );
1388                                         }
1389                                         $num = substr( $iranian[0], -2 );
1390                                         break;
1391                                 case 'xit':
1392                                         $usedIranianYear = true;
1393                                         if ( !$iranian ) {
1394                                                 $iranian = self::tsToIranian( $ts );
1395                                         }
1396                                         $num = self::$IRANIAN_DAYS[$iranian[1] - 1];
1397                                         break;
1398                                 case 'xiz':
1399                                         $usedIranianYear = true;
1400                                         if ( !$iranian ) {
1401                                                 $iranian = self::tsToIranian( $ts );
1402                                         }
1403                                         $num = $iranian[3];
1404                                         break;
1405                                 case 'a':
1406                                         $usedAMPM = true;
1407                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'am' : 'pm';
1408                                         break;
1409                                 case 'A':
1410                                         $usedAMPM = true;
1411                                         $s .= intval( substr( $ts, 8, 2 ) ) < 12 ? 'AM' : 'PM';
1412                                         break;
1413                                 case 'g':
1414                                         $usedHour = true;
1415                                         $h = substr( $ts, 8, 2 );
1416                                         $num = $h % 12 ?: 12;
1417                                         break;
1418                                 case 'G':
1419                                         $usedHour = true;
1420                                         $num = intval( substr( $ts, 8, 2 ) );
1421                                         break;
1422                                 case 'h':
1423                                         $usedHour = true;
1424                                         $h = substr( $ts, 8, 2 );
1425                                         $num = sprintf( '%02d', $h % 12 ?: 12 );
1426                                         break;
1427                                 case 'H':
1428                                         $usedHour = true;
1429                                         $num = substr( $ts, 8, 2 );
1430                                         break;
1431                                 case 'i':
1432                                         $usedMinute = true;
1433                                         $num = substr( $ts, 10, 2 );
1434                                         break;
1435                                 case 's':
1436                                         $usedSecond = true;
1437                                         $num = substr( $ts, 12, 2 );
1438                                         break;
1439                                 case 'c':
1440                                 case 'r':
1441                                         $usedSecond = true;
1442                                         // fall through
1443                                 case 'e':
1444                                 case 'O':
1445                                 case 'P':
1446                                 case 'T':
1447                                         $s .= self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1448                                         break;
1449                                 case 'w':
1450                                 case 'N':
1451                                 case 'z':
1452                                         $usedDay = true;
1453                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1454                                         break;
1455                                 case 'W':
1456                                         $usedWeek = true;
1457                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1458                                         break;
1459                                 case 't':
1460                                         $usedMonth = true;
1461                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1462                                         break;
1463                                 case 'L':
1464                                         $usedIsLeapYear = true;
1465                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1466                                         break;
1467                                 case 'o':
1468                                         $usedISOYear = true;
1469                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1470                                         break;
1471                                 case 'U':
1472                                         $usedSecond = true;
1473                                         // fall through
1474                                 case 'I':
1475                                 case 'Z':
1476                                         $num = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, $code );
1477                                         break;
1478                                 case '\\':
1479                                         # Backslash escaping
1480                                         if ( $p < $formatLength - 1 ) {
1481                                                 $s .= $format[++$p];
1482                                         } else {
1483                                                 $s .= '\\';
1484                                         }
1485                                         break;
1486                                 case '"':
1487                                         # Quoted literal
1488                                         if ( $p < $formatLength - 1 ) {
1489                                                 $endQuote = strpos( $format, '"', $p + 1 );
1490                                                 if ( $endQuote === false ) {
1491                                                         # No terminating quote, assume literal "
1492                                                         $s .= '"';
1493                                                 } else {
1494                                                         $s .= substr( $format, $p + 1, $endQuote - $p - 1 );
1495                                                         $p = $endQuote;
1496                                                 }
1497                                         } else {
1498                                                 # Quote at end of string, assume literal "
1499                                                 $s .= '"';
1500                                         }
1501                                         break;
1502                                 default:
1503                                         $s .= $format[$p];
1504                         }
1505                         if ( $num !== false ) {
1506                                 if ( $rawToggle || $raw ) {
1507                                         $s .= $num;
1508                                         $raw = false;
1509                                 } elseif ( $roman ) {
1510                                         $s .= self::romanNumeral( $num );
1511                                         $roman = false;
1512                                 } elseif ( $hebrewNum ) {
1513                                         $s .= self::hebrewNumeral( $num );
1514                                         $hebrewNum = false;
1515                                 } else {
1516                                         $s .= $this->formatNum( $num, true );
1517                                 }
1518                         }
1519                 }
1520
1521                 if ( $ttl === 'unused' ) {
1522                         // No need to calculate the TTL, the caller wont use it anyway.
1523                 } elseif ( $usedSecond ) {
1524                         $ttl = 1;
1525                 } elseif ( $usedMinute ) {
1526                         $ttl = 60 - substr( $ts, 12, 2 );
1527                 } elseif ( $usedHour ) {
1528                         $ttl = 3600 - substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1529                 } elseif ( $usedAMPM ) {
1530                         $ttl = 43200 - ( substr( $ts, 8, 2 ) % 12 ) * 3600 -
1531                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1532                 } elseif (
1533                         $usedDay ||
1534                         $usedHebrewMonth ||
1535                         $usedIranianMonth ||
1536                         $usedHijriMonth ||
1537                         $usedHebrewYear ||
1538                         $usedIranianYear ||
1539                         $usedHijriYear ||
1540                         $usedTennoYear
1541                 ) {
1542                         // @todo Someone who understands the non-Gregorian calendars
1543                         // should write proper logic for them so that they don't need purged every day.
1544                         $ttl = 86400 - substr( $ts, 8, 2 ) * 3600 -
1545                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1546                 } else {
1547                         $possibleTtls = [];
1548                         $timeRemainingInDay = 86400 - substr( $ts, 8, 2 ) * 3600 -
1549                                 substr( $ts, 10, 2 ) * 60 - substr( $ts, 12, 2 );
1550                         if ( $usedWeek ) {
1551                                 $possibleTtls[] =
1552                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400 +
1553                                         $timeRemainingInDay;
1554                         } elseif ( $usedISOYear ) {
1555                                 // December 28th falls on the last ISO week of the year, every year.
1556                                 // The last ISO week of a year can be 52 or 53.
1557                                 $lastWeekOfISOYear = DateTime::createFromFormat(
1558                                         'Ymd',
1559                                         substr( $ts, 0, 4 ) . '1228',
1560                                         $zone ?: new DateTimeZone( 'UTC' )
1561                                 )->format( 'W' );
1562                                 $currentISOWeek = self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'W' );
1563                                 $weeksRemaining = $lastWeekOfISOYear - $currentISOWeek;
1564                                 $timeRemainingInWeek =
1565                                         ( 7 - self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'N' ) ) * 86400
1566                                         + $timeRemainingInDay;
1567                                 $possibleTtls[] = $weeksRemaining * 604800 + $timeRemainingInWeek;
1568                         }
1569
1570                         if ( $usedMonth ) {
1571                                 $possibleTtls[] =
1572                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 't' ) -
1573                                                 substr( $ts, 6, 2 ) ) * 86400
1574                                         + $timeRemainingInDay;
1575                         } elseif ( $usedYear ) {
1576                                 $possibleTtls[] =
1577                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1578                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1579                                         + $timeRemainingInDay;
1580                         } elseif ( $usedIsLeapYear ) {
1581                                 $year = substr( $ts, 0, 4 );
1582                                 $timeRemainingInYear =
1583                                         ( self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'L' ) + 364 -
1584                                                 self::dateTimeObjFormat( $dateTimeObj, $ts, $zone, 'z' ) ) * 86400
1585                                         + $timeRemainingInDay;
1586                                 $mod = $year % 4;
1587                                 if ( $mod || ( !( $year % 100 ) && $year % 400 ) ) {
1588                                         // this isn't a leap year. see when the next one starts
1589                                         $nextCandidate = $year - $mod + 4;
1590                                         if ( $nextCandidate % 100 || !( $nextCandidate % 400 ) ) {
1591                                                 $possibleTtls[] = ( $nextCandidate - $year - 1 ) * 365 * 86400 +
1592                                                         $timeRemainingInYear;
1593                                         } else {
1594                                                 $possibleTtls[] = ( $nextCandidate - $year + 3 ) * 365 * 86400 +
1595                                                         $timeRemainingInYear;
1596                                         }
1597                                 } else {
1598                                         // this is a leap year, so the next year isn't
1599                                         $possibleTtls[] = $timeRemainingInYear;
1600                                 }
1601                         }
1602
1603                         if ( $possibleTtls ) {
1604                                 $ttl = min( $possibleTtls );
1605                         }
1606                 }
1607
1608                 return $s;
1609         }
1610
1611         private static $GREG_DAYS = [ 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31 ];
1612         private static $IRANIAN_DAYS = [ 31, 31, 31, 31, 31, 31, 30, 30, 30, 30, 30, 29 ];
1613
1614         /**
1615          * Algorithm by Roozbeh Pournader and Mohammad Toossi to convert
1616          * Gregorian dates to Iranian dates. Originally written in C, it
1617          * is released under the terms of GNU Lesser General Public
1618          * License. Conversion to PHP was performed by Niklas Laxström.
1619          *
1620          * Link: http://www.farsiweb.info/jalali/jalali.c
1621          *
1622          * @param string $ts
1623          *
1624          * @return int[]
1625          */
1626         private static function tsToIranian( $ts ) {
1627                 $gy = substr( $ts, 0, 4 ) - 1600;
1628                 $gm = substr( $ts, 4, 2 ) - 1;
1629                 $gd = substr( $ts, 6, 2 ) - 1;
1630
1631                 # Days passed from the beginning (including leap years)
1632                 $gDayNo = 365 * $gy
1633                         + floor( ( $gy + 3 ) / 4 )
1634                         - floor( ( $gy + 99 ) / 100 )
1635                         + floor( ( $gy + 399 ) / 400 );
1636
1637                 // Add days of the past months of this year
1638                 for ( $i = 0; $i < $gm; $i++ ) {
1639                         $gDayNo += self::$GREG_DAYS[$i];
1640                 }
1641
1642                 // Leap years
1643                 if ( $gm > 1 && ( ( $gy % 4 === 0 && $gy % 100 !== 0 || ( $gy % 400 == 0 ) ) ) ) {
1644                         $gDayNo++;
1645                 }
1646
1647                 // Days passed in current month
1648                 $gDayNo += (int)$gd;
1649
1650                 $jDayNo = $gDayNo - 79;
1651
1652                 $jNp = floor( $jDayNo / 12053 );
1653                 $jDayNo %= 12053;
1654
1655                 $jy = 979 + 33 * $jNp + 4 * floor( $jDayNo / 1461 );
1656                 $jDayNo %= 1461;
1657
1658                 if ( $jDayNo >= 366 ) {
1659                         $jy += floor( ( $jDayNo - 1 ) / 365 );
1660                         $jDayNo = floor( ( $jDayNo - 1 ) % 365 );
1661                 }
1662
1663                 $jz = $jDayNo;
1664
1665                 for ( $i = 0; $i < 11 && $jDayNo >= self::$IRANIAN_DAYS[$i]; $i++ ) {
1666                         $jDayNo -= self::$IRANIAN_DAYS[$i];
1667                 }
1668
1669                 $jm = $i + 1;
1670                 $jd = $jDayNo + 1;
1671
1672                 return [ $jy, $jm, $jd, $jz ];
1673         }
1674
1675         /**
1676          * Converting Gregorian dates to Hijri dates.
1677          *
1678          * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
1679          *
1680          * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
1681          *
1682          * @param string $ts
1683          *
1684          * @return int[]
1685          */
1686         private static function tsToHijri( $ts ) {
1687                 $year = substr( $ts, 0, 4 );
1688                 $month = substr( $ts, 4, 2 );
1689                 $day = substr( $ts, 6, 2 );
1690
1691                 $zyr = $year;
1692                 $zd = $day;
1693                 $zm = $month;
1694                 $zy = $zyr;
1695
1696                 if (
1697                         ( $zy > 1582 ) || ( ( $zy == 1582 ) && ( $zm > 10 ) ) ||
1698                         ( ( $zy == 1582 ) && ( $zm == 10 ) && ( $zd > 14 ) )
1699                 ) {
1700                         $zjd = (int)( ( 1461 * ( $zy + 4800 + (int)( ( $zm - 14 ) / 12 ) ) ) / 4 ) +
1701                                         (int)( ( 367 * ( $zm - 2 - 12 * ( (int)( ( $zm - 14 ) / 12 ) ) ) ) / 12 ) -
1702                                         (int)( ( 3 * (int)( ( ( $zy + 4900 + (int)( ( $zm - 14 ) / 12 ) ) / 100 ) ) ) / 4 ) +
1703                                         $zd - 32075;
1704                 } else {
1705                         $zjd = 367 * $zy - (int)( ( 7 * ( $zy + 5001 + (int)( ( $zm - 9 ) / 7 ) ) ) / 4 ) +
1706                                                                 (int)( ( 275 * $zm ) / 9 ) + $zd + 1729777;
1707                 }
1708
1709                 $zl = $zjd - 1948440 + 10632;
1710                 $zn = (int)( ( $zl - 1 ) / 10631 );
1711                 $zl = $zl - 10631 * $zn + 354;
1712                 $zj = ( (int)( ( 10985 - $zl ) / 5316 ) ) * ( (int)( ( 50 * $zl ) / 17719 ) ) +
1713                         ( (int)( $zl / 5670 ) ) * ( (int)( ( 43 * $zl ) / 15238 ) );
1714                 $zl = $zl - ( (int)( ( 30 - $zj ) / 15 ) ) * ( (int)( ( 17719 * $zj ) / 50 ) ) -
1715                         ( (int)( $zj / 16 ) ) * ( (int)( ( 15238 * $zj ) / 43 ) ) + 29;
1716                 $zm = (int)( ( 24 * $zl ) / 709 );
1717                 $zd = $zl - (int)( ( 709 * $zm ) / 24 );
1718                 $zy = 30 * $zn + $zj - 30;
1719
1720                 return [ $zy, $zm, $zd ];
1721         }
1722
1723         /**
1724          * Converting Gregorian dates to Hebrew dates.
1725          *
1726          * Based on a JavaScript code by Abu Mami and Yisrael Hersch
1727          * (abu-mami@kaluach.net, http://www.kaluach.net), who permitted
1728          * to translate the relevant functions into PHP and release them under
1729          * GNU GPL.
1730          *
1731          * The months are counted from Tishrei = 1. In a leap year, Adar I is 13
1732          * and Adar II is 14. In a non-leap year, Adar is 6.
1733          *
1734          * @param string $ts
1735          *
1736          * @return int[]
1737          */
1738         private static function tsToHebrew( $ts ) {
1739                 # Parse date
1740                 $year = substr( $ts, 0, 4 );
1741                 $month = substr( $ts, 4, 2 );
1742                 $day = substr( $ts, 6, 2 );
1743
1744                 # Calculate Hebrew year
1745                 $hebrewYear = $year + 3760;
1746
1747                 # Month number when September = 1, August = 12
1748                 $month += 4;
1749                 if ( $month > 12 ) {
1750                         # Next year
1751                         $month -= 12;
1752                         $year++;
1753                         $hebrewYear++;
1754                 }
1755
1756                 # Calculate day of year from 1 September
1757                 $dayOfYear = $day;
1758                 for ( $i = 1; $i < $month; $i++ ) {
1759                         if ( $i == 6 ) {
1760                                 # February
1761                                 $dayOfYear += 28;
1762                                 # Check if the year is leap
1763                                 if ( $year % 400 == 0 || ( $year % 4 == 0 && $year % 100 > 0 ) ) {
1764                                         $dayOfYear++;
1765                                 }
1766                         } elseif ( $i == 8 || $i == 10 || $i == 1 || $i == 3 ) {
1767                                 $dayOfYear += 30;
1768                         } else {
1769                                 $dayOfYear += 31;
1770                         }
1771                 }
1772
1773                 # Calculate the start of the Hebrew year
1774                 $start = self::hebrewYearStart( $hebrewYear );
1775
1776                 # Calculate next year's start
1777                 if ( $dayOfYear <= $start ) {
1778                         # Day is before the start of the year - it is the previous year
1779                         # Next year's start
1780                         $nextStart = $start;
1781                         # Previous year
1782                         $year--;
1783                         $hebrewYear--;
1784                         # Add days since previous year's 1 September
1785                         $dayOfYear += 365;
1786                         if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1787                                 # Leap year
1788                                 $dayOfYear++;
1789                         }
1790                         # Start of the new (previous) year
1791                         $start = self::hebrewYearStart( $hebrewYear );
1792                 } else {
1793                         # Next year's start
1794                         $nextStart = self::hebrewYearStart( $hebrewYear + 1 );
1795                 }
1796
1797                 # Calculate Hebrew day of year
1798                 $hebrewDayOfYear = $dayOfYear - $start;
1799
1800                 # Difference between year's days
1801                 $diff = $nextStart - $start;
1802                 # Add 12 (or 13 for leap years) days to ignore the difference between
1803                 # Hebrew and Gregorian year (353 at least vs. 365/6) - now the
1804                 # difference is only about the year type
1805                 if ( ( $year % 400 == 0 ) || ( $year % 100 != 0 && $year % 4 == 0 ) ) {
1806                         $diff += 13;
1807                 } else {
1808                         $diff += 12;
1809                 }
1810
1811                 # Check the year pattern, and is leap year
1812                 # 0 means an incomplete year, 1 means a regular year, 2 means a complete year
1813                 # This is mod 30, to work on both leap years (which add 30 days of Adar I)
1814                 # and non-leap years
1815                 $yearPattern = $diff % 30;
1816                 # Check if leap year
1817                 $isLeap = $diff >= 30;
1818
1819                 # Calculate day in the month from number of day in the Hebrew year
1820                 # Don't check Adar - if the day is not in Adar, we will stop before;
1821                 # if it is in Adar, we will use it to check if it is Adar I or Adar II
1822                 $hebrewDay = $hebrewDayOfYear;
1823                 $hebrewMonth = 1;
1824                 $days = 0;
1825                 while ( $hebrewMonth <= 12 ) {
1826                         # Calculate days in this month
1827                         if ( $isLeap && $hebrewMonth == 6 ) {
1828                                 # Leap year - has Adar I, with 30 days, and Adar II, with 29 days
1829                                 $days = 30;
1830                                 if ( $hebrewDay <= $days ) {
1831                                         # Day in Adar I
1832                                         $hebrewMonth = 13;
1833                                 } else {
1834                                         # Subtract the days of Adar I
1835                                         $hebrewDay -= $days;
1836                                         # Try Adar II
1837                                         $days = 29;
1838                                         if ( $hebrewDay <= $days ) {
1839                                                 # Day in Adar II
1840                                                 $hebrewMonth = 14;
1841                                         }
1842                                 }
1843                         } elseif ( $hebrewMonth == 2 && $yearPattern == 2 ) {
1844                                 # Cheshvan in a complete year (otherwise as the rule below)
1845                                 $days = 30;
1846                         } elseif ( $hebrewMonth == 3 && $yearPattern == 0 ) {
1847                                 # Kislev in an incomplete year (otherwise as the rule below)
1848                                 $days = 29;
1849                         } else {
1850                                 # Odd months have 30 days, even have 29
1851                                 $days = 30 - ( $hebrewMonth - 1 ) % 2;
1852                         }
1853                         if ( $hebrewDay <= $days ) {
1854                                 # In the current month
1855                                 break;
1856                         } else {
1857                                 # Subtract the days of the current month
1858                                 $hebrewDay -= $days;
1859                                 # Try in the next month
1860                                 $hebrewMonth++;
1861                         }
1862                 }
1863
1864                 return [ $hebrewYear, $hebrewMonth, $hebrewDay, $days ];
1865         }
1866
1867         /**
1868          * This calculates the Hebrew year start, as days since 1 September.
1869          * Based on Carl Friedrich Gauss algorithm for finding Easter date.
1870          * Used for Hebrew date.
1871          *
1872          * @param int $year
1873          *
1874          * @return string
1875          */
1876         private static function hebrewYearStart( $year ) {
1877                 $a = intval( ( 12 * ( $year - 1 ) + 17 ) % 19 );
1878                 $b = intval( ( $year - 1 ) % 4 );
1879                 $m = 32.044093161144 + 1.5542417966212 * $a + $b / 4.0 - 0.0031777940220923 * ( $year - 1 );
1880                 if ( $m < 0 ) {
1881                         $m--;
1882                 }
1883                 $Mar = intval( $m );
1884                 if ( $m < 0 ) {
1885                         $m++;
1886                 }
1887                 $m -= $Mar;
1888
1889                 $c = intval( ( $Mar + 3 * ( $year - 1 ) + 5 * $b + 5 ) % 7 );
1890                 if ( $c == 0 && $a > 11 && $m >= 0.89772376543210 ) {
1891                         $Mar++;
1892                 } elseif ( $c == 1 && $a > 6 && $m >= 0.63287037037037 ) {
1893                         $Mar += 2;
1894                 } elseif ( $c == 2 || $c == 4 || $c == 6 ) {
1895                         $Mar++;
1896                 }
1897
1898                 $Mar += intval( ( $year - 3761 ) / 100 ) - intval( ( $year - 3761 ) / 400 ) - 24;
1899                 return $Mar;
1900         }
1901
1902         /**
1903          * Algorithm to convert Gregorian dates to Thai solar dates,
1904          * Minguo dates or Minguo dates.
1905          *
1906          * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
1907          *       https://en.wikipedia.org/wiki/Minguo_calendar
1908          *       https://en.wikipedia.org/wiki/Japanese_era_name
1909          *
1910          * @param string $ts 14-character timestamp
1911          * @param string $cName Calender name
1912          * @return array Converted year, month, day
1913          */
1914         private static function tsToYear( $ts, $cName ) {
1915                 $gy = substr( $ts, 0, 4 );
1916                 $gm = substr( $ts, 4, 2 );
1917                 $gd = substr( $ts, 6, 2 );
1918
1919                 if ( !strcmp( $cName, 'thai' ) ) {
1920                         # Thai solar dates
1921                         # Add 543 years to the Gregorian calendar
1922                         # Months and days are identical
1923                         $gy_offset = $gy + 543;
1924                         # fix for dates between 1912 and 1941
1925                         # https://en.wikipedia.org/?oldid=836596673#New_year
1926                         if ( $gy >= 1912 && $gy <= 1940 ) {
1927                                 if ( $gm <= 3 ) {
1928                                         $gy_offset--;
1929                                 }
1930                                 $gm = ( $gm - 3 ) % 12;
1931                         }
1932                 } elseif ( ( !strcmp( $cName, 'minguo' ) ) || !strcmp( $cName, 'juche' ) ) {
1933                         # Minguo dates
1934                         # Deduct 1911 years from the Gregorian calendar
1935                         # Months and days are identical
1936                         $gy_offset = $gy - 1911;
1937                 } elseif ( !strcmp( $cName, 'tenno' ) ) {
1938                         # Nengō dates up to Meiji period
1939                         # Deduct years from the Gregorian calendar
1940                         # depending on the nengo periods
1941                         # Months and days are identical
1942                         if ( ( $gy < 1912 )
1943                                 || ( ( $gy == 1912 ) && ( $gm < 7 ) )
1944                                 || ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd < 31 ) )
1945                         ) {
1946                                 # Meiji period
1947                                 $gy_gannen = $gy - 1868 + 1;
1948                                 $gy_offset = $gy_gannen;
1949                                 if ( $gy_gannen == 1 ) {
1950                                         $gy_offset = '元';
1951                                 }
1952                                 $gy_offset = '明治' . $gy_offset;
1953                         } elseif (
1954                                 ( ( $gy == 1912 ) && ( $gm == 7 ) && ( $gd == 31 ) ) ||
1955                                 ( ( $gy == 1912 ) && ( $gm >= 8 ) ) ||
1956                                 ( ( $gy > 1912 ) && ( $gy < 1926 ) ) ||
1957                                 ( ( $gy == 1926 ) && ( $gm < 12 ) ) ||
1958                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd < 26 ) )
1959                         ) {
1960                                 # Taishō period
1961                                 $gy_gannen = $gy - 1912 + 1;
1962                                 $gy_offset = $gy_gannen;
1963                                 if ( $gy_gannen == 1 ) {
1964                                         $gy_offset = '元';
1965                                 }
1966                                 $gy_offset = '大正' . $gy_offset;
1967                         } elseif (
1968                                 ( ( $gy == 1926 ) && ( $gm == 12 ) && ( $gd >= 26 ) ) ||
1969                                 ( ( $gy > 1926 ) && ( $gy < 1989 ) ) ||
1970                                 ( ( $gy == 1989 ) && ( $gm == 1 ) && ( $gd < 8 ) )
1971                         ) {
1972                                 # Shōwa period
1973                                 $gy_gannen = $gy - 1926 + 1;
1974                                 $gy_offset = $gy_gannen;
1975                                 if ( $gy_gannen == 1 ) {
1976                                         $gy_offset = '元';
1977                                 }
1978                                 $gy_offset = '昭和' . $gy_offset;
1979                         } else {
1980                                 # Heisei period
1981                                 $gy_gannen = $gy - 1989 + 1;
1982                                 $gy_offset = $gy_gannen;
1983                                 if ( $gy_gannen == 1 ) {
1984                                         $gy_offset = '元';
1985                                 }
1986                                 $gy_offset = '平成' . $gy_offset;
1987                         }
1988                 } else {
1989                         $gy_offset = $gy;
1990                 }
1991
1992                 return [ $gy_offset, $gm, $gd ];
1993         }
1994
1995         /**
1996          * Gets directionality of the first strongly directional codepoint, for embedBidi()
1997          *
1998          * This is the rule the BIDI algorithm uses to determine the directionality of
1999          * paragraphs ( https://www.unicode.org/reports/tr9/#The_Paragraph_Level ) and
2000          * FSI isolates ( https://www.unicode.org/reports/tr9/#Explicit_Directional_Isolates ).
2001          *
2002          * TODO: Does not handle BIDI control characters inside the text.
2003          * TODO: Does not handle unallocated characters.
2004          *
2005          * @param string $text Text to test
2006          * @return null|string Directionality ('ltr' or 'rtl') or null
2007          */
2008         private static function strongDirFromContent( $text = '' ) {
2009                 if ( !preg_match( self::$strongDirRegex, $text, $matches ) ) {
2010                         return null;
2011                 }
2012                 if ( $matches[1] === '' ) {
2013                         return 'rtl';
2014                 }
2015                 return 'ltr';
2016         }
2017
2018         /**
2019          * Roman number formatting up to 10000
2020          *
2021          * @param int $num
2022          *
2023          * @return string
2024          */
2025         static function romanNumeral( $num ) {
2026                 static $table = [
2027                         [ '', 'I', 'II', 'III', 'IV', 'V', 'VI', 'VII', 'VIII', 'IX', 'X' ],
2028                         [ '', 'X', 'XX', 'XXX', 'XL', 'L', 'LX', 'LXX', 'LXXX', 'XC', 'C' ],
2029                         [ '', 'C', 'CC', 'CCC', 'CD', 'D', 'DC', 'DCC', 'DCCC', 'CM', 'M' ],
2030                         [ '', 'M', 'MM', 'MMM', 'MMMM', 'MMMMM', 'MMMMMM', 'MMMMMMM',
2031                                 'MMMMMMMM', 'MMMMMMMMM', 'MMMMMMMMMM' ]
2032                 ];
2033
2034                 $num = intval( $num );
2035                 if ( $num > 10000 || $num <= 0 ) {
2036                         return $num;
2037                 }
2038
2039                 $s = '';
2040                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2041                         if ( $num >= $pow10 ) {
2042                                 $s .= $table[$i][(int)floor( $num / $pow10 )];
2043                         }
2044                         $num = $num % $pow10;
2045                 }
2046                 return $s;
2047         }
2048
2049         /**
2050          * Hebrew Gematria number formatting up to 9999
2051          *
2052          * @param int $num
2053          *
2054          * @return string
2055          */
2056         static function hebrewNumeral( $num ) {
2057                 static $table = [
2058                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ],
2059                         [ '', 'י', 'כ', 'ל', 'מ', 'נ', 'ס', 'ע', 'פ', 'צ', 'ק' ],
2060                         [ '',
2061                                 [ 'ק' ],
2062                                 [ 'ר' ],
2063                                 [ 'ש' ],
2064                                 [ 'ת' ],
2065                                 [ 'ת', 'ק' ],
2066                                 [ 'ת', 'ר' ],
2067                                 [ 'ת', 'ש' ],
2068                                 [ 'ת', 'ת' ],
2069                                 [ 'ת', 'ת', 'ק' ],
2070                                 [ 'ת', 'ת', 'ר' ],
2071                         ],
2072                         [ '', 'א', 'ב', 'ג', 'ד', 'ה', 'ו', 'ז', 'ח', 'ט', 'י' ]
2073                 ];
2074
2075                 $num = intval( $num );
2076                 if ( $num > 9999 || $num <= 0 ) {
2077                         return $num;
2078                 }
2079
2080                 // Round thousands have special notations
2081                 if ( $num === 1000 ) {
2082                         return "א' אלף";
2083                 } elseif ( $num % 1000 === 0 ) {
2084                         return $table[0][$num / 1000] . "' אלפים";
2085                 }
2086
2087                 $letters = [];
2088
2089                 for ( $pow10 = 1000, $i = 3; $i >= 0; $pow10 /= 10, $i-- ) {
2090                         if ( $num >= $pow10 ) {
2091                                 if ( $num === 15 || $num === 16 ) {
2092                                         $letters[] = $table[0][9];
2093                                         $letters[] = $table[0][$num - 9];
2094                                         $num = 0;
2095                                 } else {
2096                                         $letters = array_merge(
2097                                                 $letters,
2098                                                 (array)$table[$i][intval( $num / $pow10 )]
2099                                         );
2100
2101                                         if ( $pow10 === 1000 ) {
2102                                                 $letters[] = "'";
2103                                         }
2104                                 }
2105                         }
2106
2107                         $num = $num % $pow10;
2108                 }
2109
2110                 $preTransformLength = count( $letters );
2111                 if ( $preTransformLength === 1 ) {
2112                         // Add geresh (single quote) to one-letter numbers
2113                         $letters[] = "'";
2114                 } else {
2115                         $lastIndex = $preTransformLength - 1;
2116                         $letters[$lastIndex] = str_replace(
2117                                 [ 'כ', 'מ', 'נ', 'פ', 'צ' ],
2118                                 [ 'ך', 'ם', 'ן', 'ף', 'ץ' ],
2119                                 $letters[$lastIndex]
2120                         );
2121
2122                         // Add gershayim (double quote) to multiple-letter numbers,
2123                         // but exclude numbers with only one letter after the thousands
2124                         // (1001-1009, 1020, 1030, 2001-2009, etc.)
2125                         if ( $letters[1] === "'" && $preTransformLength === 3 ) {
2126                                 $letters[] = "'";
2127                         } else {
2128                                 array_splice( $letters, -1, 0, '"' );
2129                         }
2130                 }
2131
2132                 return implode( $letters );
2133         }
2134
2135         /**
2136          * Used by date() and time() to adjust the time output.
2137          *
2138          * @param string $ts The time in date('YmdHis') format
2139          * @param mixed $tz Adjust the time by this amount (default false, mean we
2140          *   get user timecorrection setting)
2141          * @return int
2142          */
2143         public function userAdjust( $ts, $tz = false ) {
2144                 global $wgUser, $wgLocalTZoffset;
2145
2146                 if ( $tz === false ) {
2147                         $tz = $wgUser->getOption( 'timecorrection' );
2148                 }
2149
2150                 $data = explode( '|', $tz, 3 );
2151
2152                 if ( $data[0] == 'ZoneInfo' ) {
2153                         try {
2154                                 $userTZ = new DateTimeZone( $data[2] );
2155                                 $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
2156                                 $date->setTimezone( $userTZ );
2157                                 return $date->format( 'YmdHis' );
2158                         } catch ( Exception $e ) {
2159                                 // Unrecognized timezone, default to 'Offset' with the stored offset.
2160                                 $data[0] = 'Offset';
2161                         }
2162                 }
2163
2164                 if ( $data[0] == 'System' || $tz == '' ) {
2165                         # Global offset in minutes.
2166                         $minDiff = $wgLocalTZoffset;
2167                 } elseif ( $data[0] == 'Offset' ) {
2168                         $minDiff = intval( $data[1] );
2169                 } else {
2170                         $data = explode( ':', $tz );
2171                         if ( count( $data ) == 2 ) {
2172                                 $data[0] = intval( $data[0] );
2173                                 $data[1] = intval( $data[1] );
2174                                 $minDiff = abs( $data[0] ) * 60 + $data[1];
2175                                 if ( $data[0] < 0 ) {
2176                                         $minDiff = -$minDiff;
2177                                 }
2178                         } else {
2179                                 $minDiff = intval( $data[0] ) * 60;
2180                         }
2181                 }
2182
2183                 # No difference ? Return time unchanged
2184                 if ( $minDiff == 0 ) {
2185                         return $ts;
2186                 }
2187
2188                 Wikimedia\suppressWarnings(); // E_STRICT system time bitching
2189                 # Generate an adjusted date; take advantage of the fact that mktime
2190                 # will normalize out-of-range values so we don't have to split $minDiff
2191                 # into hours and minutes.
2192                 $t = mktime( (
2193                         (int)substr( $ts, 8, 2 ) ), # Hours
2194                         (int)substr( $ts, 10, 2 ) + $minDiff, # Minutes
2195                         (int)substr( $ts, 12, 2 ), # Seconds
2196                         (int)substr( $ts, 4, 2 ), # Month
2197                         (int)substr( $ts, 6, 2 ), # Day
2198                         (int)substr( $ts, 0, 4 ) ); # Year
2199
2200                 $date = date( 'YmdHis', $t );
2201                 Wikimedia\restoreWarnings();
2202
2203                 return $date;
2204         }
2205
2206         /**
2207          * This is meant to be used by time(), date(), and timeanddate() to get
2208          * the date preference they're supposed to use, it should be used in
2209          * all children.
2210          *
2211          *     function timeanddate([...], $format = true) {
2212          *       $datePreference = $this->dateFormat($format);
2213          *       [...]
2214          *     }
2215          *
2216          * @param int|string|bool $usePrefs If true, the user's preference is used
2217          *   if false, the site/language default is used
2218          *   if int/string, assumed to be a format.
2219          * @return string
2220          */
2221         function dateFormat( $usePrefs = true ) {
2222                 global $wgUser;
2223
2224                 if ( is_bool( $usePrefs ) ) {
2225                         if ( $usePrefs ) {
2226                                 $datePreference = $wgUser->getDatePreference();
2227                         } else {
2228                                 $datePreference = (string)User::getDefaultOption( 'date' );
2229                         }
2230                 } else {
2231                         $datePreference = (string)$usePrefs;
2232                 }
2233
2234                 // return int
2235                 if ( $datePreference == '' ) {
2236                         return 'default';
2237                 }
2238
2239                 return $datePreference;
2240         }
2241
2242         /**
2243          * Get a format string for a given type and preference
2244          * @param string $type May be 'date', 'time', 'both', or 'pretty'.
2245          * @param string $pref The format name as it appears in Messages*.php under
2246          *  $datePreferences.
2247          *
2248          * @since 1.22 New type 'pretty' that provides a more readable timestamp format
2249          *
2250          * @return string
2251          */
2252         function getDateFormatString( $type, $pref ) {
2253                 $wasDefault = false;
2254                 if ( $pref == 'default' ) {
2255                         $wasDefault = true;
2256                         $pref = $this->getDefaultDateFormat();
2257                 }
2258
2259                 if ( !isset( $this->dateFormatStrings[$type][$pref] ) ) {
2260                         $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2261
2262                         if ( $type === 'pretty' && $df === null ) {
2263                                 $df = $this->getDateFormatString( 'date', $pref );
2264                         }
2265
2266                         if ( !$wasDefault && $df === null ) {
2267                                 $pref = $this->getDefaultDateFormat();
2268                                 $df = self::$dataCache->getSubitem( $this->mCode, 'dateFormats', "$pref $type" );
2269                         }
2270
2271                         $this->dateFormatStrings[$type][$pref] = $df;
2272                 }
2273                 return $this->dateFormatStrings[$type][$pref];
2274         }
2275
2276         /**
2277          * @param string $ts The time format which needs to be turned into a
2278          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2279          * @param bool $adj Whether to adjust the time output according to the
2280          *   user configured offset ($timecorrection)
2281          * @param mixed $format True to use user's date format preference
2282          * @param string|bool $timecorrection The time offset as returned by
2283          *   validateTimeZone() in Special:Preferences
2284          * @return string
2285          */
2286         public function date( $ts, $adj = false, $format = true, $timecorrection = false ) {
2287                 $ts = wfTimestamp( TS_MW, $ts );
2288                 if ( $adj ) {
2289                         $ts = $this->userAdjust( $ts, $timecorrection );
2290                 }
2291                 $df = $this->getDateFormatString( 'date', $this->dateFormat( $format ) );
2292                 return $this->sprintfDate( $df, $ts );
2293         }
2294
2295         /**
2296          * @param string $ts The time format which needs to be turned into a
2297          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2298          * @param bool $adj Whether to adjust the time output according to the
2299          *   user configured offset ($timecorrection)
2300          * @param mixed $format True to use user's date format preference
2301          * @param string|bool $timecorrection The time offset as returned by
2302          *   validateTimeZone() in Special:Preferences
2303          * @return string
2304          */
2305         public function time( $ts, $adj = false, $format = true, $timecorrection = false ) {
2306                 $ts = wfTimestamp( TS_MW, $ts );
2307                 if ( $adj ) {
2308                         $ts = $this->userAdjust( $ts, $timecorrection );
2309                 }
2310                 $df = $this->getDateFormatString( 'time', $this->dateFormat( $format ) );
2311                 return $this->sprintfDate( $df, $ts );
2312         }
2313
2314         /**
2315          * @param string $ts The time format which needs to be turned into a
2316          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2317          * @param bool $adj Whether to adjust the time output according to the
2318          *   user configured offset ($timecorrection)
2319          * @param mixed $format What format to return, if it's false output the
2320          *   default one (default true)
2321          * @param string|bool $timecorrection The time offset as returned by
2322          *   validateTimeZone() in Special:Preferences
2323          * @return string
2324          */
2325         public function timeanddate( $ts, $adj = false, $format = true, $timecorrection = false ) {
2326                 $ts = wfTimestamp( TS_MW, $ts );
2327                 if ( $adj ) {
2328                         $ts = $this->userAdjust( $ts, $timecorrection );
2329                 }
2330                 $df = $this->getDateFormatString( 'both', $this->dateFormat( $format ) );
2331                 return $this->sprintfDate( $df, $ts );
2332         }
2333
2334         /**
2335          * Takes a number of seconds and turns it into a text using values such as hours and minutes.
2336          *
2337          * @since 1.20
2338          *
2339          * @param int $seconds The amount of seconds.
2340          * @param array $chosenIntervals The intervals to enable.
2341          *
2342          * @return string
2343          */
2344         public function formatDuration( $seconds, array $chosenIntervals = [] ) {
2345                 $intervals = $this->getDurationIntervals( $seconds, $chosenIntervals );
2346
2347                 $segments = [];
2348
2349                 foreach ( $intervals as $intervalName => $intervalValue ) {
2350                         // Messages: duration-seconds, duration-minutes, duration-hours, duration-days, duration-weeks,
2351                         // duration-years, duration-decades, duration-centuries, duration-millennia
2352                         $message = wfMessage( 'duration-' . $intervalName )->numParams( $intervalValue );
2353                         $segments[] = $message->inLanguage( $this )->escaped();
2354                 }
2355
2356                 return $this->listToText( $segments );
2357         }
2358
2359         /**
2360          * Takes a number of seconds and returns an array with a set of corresponding intervals.
2361          * For example 65 will be turned into [ minutes => 1, seconds => 5 ].
2362          *
2363          * @since 1.20
2364          *
2365          * @param int $seconds The amount of seconds.
2366          * @param array $chosenIntervals The intervals to enable.
2367          *
2368          * @return array
2369          */
2370         public function getDurationIntervals( $seconds, array $chosenIntervals = [] ) {
2371                 if ( empty( $chosenIntervals ) ) {
2372                         $chosenIntervals = [
2373                                 'millennia',
2374                                 'centuries',
2375                                 'decades',
2376                                 'years',
2377                                 'days',
2378                                 'hours',
2379                                 'minutes',
2380                                 'seconds'
2381                         ];
2382                 }
2383
2384                 $intervals = array_intersect_key( self::$durationIntervals, array_flip( $chosenIntervals ) );
2385                 $sortedNames = array_keys( $intervals );
2386                 $smallestInterval = array_pop( $sortedNames );
2387
2388                 $segments = [];
2389
2390                 foreach ( $intervals as $name => $length ) {
2391                         $value = floor( $seconds / $length );
2392
2393                         if ( $value > 0 || ( $name == $smallestInterval && empty( $segments ) ) ) {
2394                                 $seconds -= $value * $length;
2395                                 $segments[$name] = $value;
2396                         }
2397                 }
2398
2399                 return $segments;
2400         }
2401
2402         /**
2403          * Internal helper function for userDate(), userTime() and userTimeAndDate()
2404          *
2405          * @param string $type Can be 'date', 'time' or 'both'
2406          * @param string $ts The time format which needs to be turned into a
2407          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2408          * @param User $user User object used to get preferences for timezone and format
2409          * @param array $options Array, can contain the following keys:
2410          *   - 'timecorrection': time correction, can have the following values:
2411          *     - true: use user's preference
2412          *     - false: don't use time correction
2413          *     - int: value of time correction in minutes
2414          *   - 'format': format to use, can have the following values:
2415          *     - true: use user's preference
2416          *     - false: use default preference
2417          *     - string: format to use
2418          * @since 1.19
2419          * @return string
2420          */
2421         private function internalUserTimeAndDate( $type, $ts, User $user, array $options ) {
2422                 $ts = wfTimestamp( TS_MW, $ts );
2423                 $options += [ 'timecorrection' => true, 'format' => true ];
2424                 if ( $options['timecorrection'] !== false ) {
2425                         if ( $options['timecorrection'] === true ) {
2426                                 $offset = $user->getOption( 'timecorrection' );
2427                         } else {
2428                                 $offset = $options['timecorrection'];
2429                         }
2430                         $ts = $this->userAdjust( $ts, $offset );
2431                 }
2432                 if ( $options['format'] === true ) {
2433                         $format = $user->getDatePreference();
2434                 } else {
2435                         $format = $options['format'];
2436                 }
2437                 $df = $this->getDateFormatString( $type, $this->dateFormat( $format ) );
2438                 return $this->sprintfDate( $df, $ts );
2439         }
2440
2441         /**
2442          * Get the formatted date for the given timestamp and formatted for
2443          * the given user.
2444          *
2445          * @param mixed $ts Mixed: the time format which needs to be turned into a
2446          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2447          * @param User $user User object used to get preferences for timezone and format
2448          * @param array $options Array, can contain the following keys:
2449          *   - 'timecorrection': time correction, can have the following values:
2450          *     - true: use user's preference
2451          *     - false: don't use time correction
2452          *     - int: value of time correction in minutes
2453          *   - 'format': format to use, can have the following values:
2454          *     - true: use user's preference
2455          *     - false: use default preference
2456          *     - string: format to use
2457          * @since 1.19
2458          * @return string
2459          */
2460         public function userDate( $ts, User $user, array $options = [] ) {
2461                 return $this->internalUserTimeAndDate( 'date', $ts, $user, $options );
2462         }
2463
2464         /**
2465          * Get the formatted time for the given timestamp and formatted for
2466          * the given user.
2467          *
2468          * @param mixed $ts The time format which needs to be turned into a
2469          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2470          * @param User $user User object used to get preferences for timezone and format
2471          * @param array $options Array, can contain the following keys:
2472          *   - 'timecorrection': time correction, can have the following values:
2473          *     - true: use user's preference
2474          *     - false: don't use time correction
2475          *     - int: value of time correction in minutes
2476          *   - 'format': format to use, can have the following values:
2477          *     - true: use user's preference
2478          *     - false: use default preference
2479          *     - string: format to use
2480          * @since 1.19
2481          * @return string
2482          */
2483         public function userTime( $ts, User $user, array $options = [] ) {
2484                 return $this->internalUserTimeAndDate( 'time', $ts, $user, $options );
2485         }
2486
2487         /**
2488          * Get the formatted date and time for the given timestamp and formatted for
2489          * the given user.
2490          *
2491          * @param mixed $ts The time format which needs to be turned into a
2492          *   date('YmdHis') format with wfTimestamp(TS_MW,$ts)
2493          * @param User $user User object used to get preferences for timezone and format
2494          * @param array $options Array, can contain the following keys:
2495          *   - 'timecorrection': time correction, can have the following values:
2496          *     - true: use user's preference
2497          *     - false: don't use time correction
2498          *     - int: value of time correction in minutes
2499          *   - 'format': format to use, can have the following values:
2500          *     - true: use user's preference
2501          *     - false: use default preference
2502          *     - string: format to use
2503          * @since 1.19
2504          * @return string
2505          */
2506         public function userTimeAndDate( $ts, User $user, array $options = [] ) {
2507                 return $this->internalUserTimeAndDate( 'both', $ts, $user, $options );
2508         }
2509
2510         /**
2511          * Get the timestamp in a human-friendly relative format, e.g., "3 days ago".
2512          *
2513          * Determine the difference between the timestamp and the current time, and
2514          * generate a readable timestamp by returning "<N> <units> ago", where the
2515          * largest possible unit is used.
2516          *
2517          * @since 1.26 (Prior to 1.26 method existed but was not meant to be used directly)
2518          *
2519          * @param MWTimestamp $time
2520          * @param MWTimestamp|null $relativeTo The base timestamp to compare to (defaults to now)
2521          * @param User|null $user User the timestamp is being generated for
2522          *  (or null to use main context's user)
2523          * @return string Formatted timestamp
2524          */
2525         public function getHumanTimestamp(
2526                 MWTimestamp $time, MWTimestamp $relativeTo = null, User $user = null
2527         ) {
2528                 if ( $relativeTo === null ) {
2529                         $relativeTo = new MWTimestamp();
2530                 }
2531                 if ( $user === null ) {
2532                         $user = RequestContext::getMain()->getUser();
2533                 }
2534
2535                 // Adjust for the user's timezone.
2536                 $offsetThis = $time->offsetForUser( $user );
2537                 $offsetRel = $relativeTo->offsetForUser( $user );
2538
2539                 $ts = '';
2540                 if ( Hooks::run( 'GetHumanTimestamp', [ &$ts, $time, $relativeTo, $user, $this ] ) ) {
2541                         $ts = $this->getHumanTimestampInternal( $time, $relativeTo, $user );
2542                 }
2543
2544                 // Reset the timezone on the objects.
2545                 $time->timestamp->sub( $offsetThis );
2546                 $relativeTo->timestamp->sub( $offsetRel );
2547
2548                 return $ts;
2549         }
2550
2551         /**
2552          * Convert an MWTimestamp into a pretty human-readable timestamp using
2553          * the given user preferences and relative base time.
2554          *
2555          * @see Language::getHumanTimestamp
2556          * @param MWTimestamp $ts Timestamp to prettify
2557          * @param MWTimestamp $relativeTo Base timestamp
2558          * @param User $user User preferences to use
2559          * @return string Human timestamp
2560          * @since 1.26
2561          */
2562         private function getHumanTimestampInternal(
2563                 MWTimestamp $ts, MWTimestamp $relativeTo, User $user
2564         ) {
2565                 $diff = $ts->diff( $relativeTo );
2566                 $diffDay = (bool)( (int)$ts->timestamp->format( 'w' ) -
2567                         (int)$relativeTo->timestamp->format( 'w' ) );
2568                 $days = $diff->days ?: (int)$diffDay;
2569                 if ( $diff->invert || $days > 5
2570                         && $ts->timestamp->format( 'Y' ) !== $relativeTo->timestamp->format( 'Y' )
2571                 ) {
2572                         // Timestamps are in different years: use full timestamp
2573                         // Also do full timestamp for future dates
2574                         /**
2575                          * @todo FIXME: Add better handling of future timestamps.
2576                          */
2577                         $format = $this->getDateFormatString( 'both', $user->getDatePreference() ?: 'default' );
2578                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2579                 } elseif ( $days > 5 ) {
2580                         // Timestamps are in same year,  but more than 5 days ago: show day and month only.
2581                         $format = $this->getDateFormatString( 'pretty', $user->getDatePreference() ?: 'default' );
2582                         $ts = $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) );
2583                 } elseif ( $days > 1 ) {
2584                         // Timestamp within the past week: show the day of the week and time
2585                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2586                         $weekday = self::$mWeekdayMsgs[$ts->timestamp->format( 'w' )];
2587                         // Messages:
2588                         // sunday-at, monday-at, tuesday-at, wednesday-at, thursday-at, friday-at, saturday-at
2589                         $ts = wfMessage( "$weekday-at" )
2590                                 ->inLanguage( $this )
2591                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2592                                 ->text();
2593                 } elseif ( $days == 1 ) {
2594                         // Timestamp was yesterday: say 'yesterday' and the time.
2595                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2596                         $ts = wfMessage( 'yesterday-at' )
2597                                 ->inLanguage( $this )
2598                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2599                                 ->text();
2600                 } elseif ( $diff->h > 1 || $diff->h == 1 && $diff->i > 30 ) {
2601                         // Timestamp was today, but more than 90 minutes ago: say 'today' and the time.
2602                         $format = $this->getDateFormatString( 'time', $user->getDatePreference() ?: 'default' );
2603                         $ts = wfMessage( 'today-at' )
2604                                 ->inLanguage( $this )
2605                                 ->params( $this->sprintfDate( $format, $ts->getTimestamp( TS_MW ) ) )
2606                                 ->text();
2607
2608                 // From here on in, the timestamp was soon enough ago so that we can simply say
2609                 // XX units ago, e.g., "2 hours ago" or "5 minutes ago"
2610                 } elseif ( $diff->h == 1 ) {
2611                         // Less than 90 minutes, but more than an hour ago.
2612                         $ts = wfMessage( 'hours-ago' )->inLanguage( $this )->numParams( 1 )->text();
2613                 } elseif ( $diff->i >= 1 ) {
2614                         // A few minutes ago.
2615                         $ts = wfMessage( 'minutes-ago' )->inLanguage( $this )->numParams( $diff->i )->text();
2616                 } elseif ( $diff->s >= 30 ) {
2617                         // Less than a minute, but more than 30 sec ago.
2618                         $ts = wfMessage( 'seconds-ago' )->inLanguage( $this )->numParams( $diff->s )->text();
2619                 } else {
2620                         // Less than 30 seconds ago.
2621                         $ts = wfMessage( 'just-now' )->text();
2622                 }
2623
2624                 return $ts;
2625         }
2626
2627         /**
2628          * @param string $key
2629          * @return string|null
2630          */
2631         public function getMessage( $key ) {
2632                 return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
2633         }
2634
2635         /**
2636          * @return array
2637          */
2638         function getAllMessages() {
2639                 return self::$dataCache->getItem( $this->mCode, 'messages' );
2640         }
2641
2642         /**
2643          * @param string $in
2644          * @param string $out
2645          * @param string $string
2646          * @return string
2647          */
2648         public function iconv( $in, $out, $string ) {
2649                 # Even with //IGNORE iconv can whine about illegal characters in
2650                 # *input* string. We just ignore those too.
2651                 # REF: https://bugs.php.net/bug.php?id=37166
2652                 # REF: https://phabricator.wikimedia.org/T18885
2653                 Wikimedia\suppressWarnings();
2654                 $text = iconv( $in, $out . '//IGNORE', $string );
2655                 Wikimedia\restoreWarnings();
2656                 return $text;
2657         }
2658
2659         // callback functions for ucwords(), ucwordbreaks()
2660
2661         /**
2662          * @param array $matches
2663          * @return mixed|string
2664          */
2665         function ucwordbreaksCallbackAscii( $matches ) {
2666                 return $this->ucfirst( $matches[1] );
2667         }
2668
2669         /**
2670          * @param array $matches
2671          * @return string
2672          */
2673         function ucwordbreaksCallbackMB( $matches ) {
2674                 return mb_strtoupper( $matches[0] );
2675         }
2676
2677         /**
2678          * @param array $matches
2679          * @return string
2680          */
2681         function ucwordsCallbackMB( $matches ) {
2682                 return mb_strtoupper( $matches[0] );
2683         }
2684
2685         /**
2686          * Make a string's first character uppercase
2687          *
2688          * @param string $str
2689          *
2690          * @return string
2691          */
2692         public function ucfirst( $str ) {
2693                 $o = ord( $str );
2694                 if ( $o < 96 ) { // if already uppercase...
2695                         return $str;
2696                 } elseif ( $o < 128 ) {
2697                         return ucfirst( $str ); // use PHP's ucfirst()
2698                 } else {
2699                         // fall back to more complex logic in case of multibyte strings
2700                         return $this->uc( $str, true );
2701                 }
2702         }
2703
2704         /**
2705          * Convert a string to uppercase
2706          *
2707          * @param string $str
2708          * @param bool $first
2709          *
2710          * @return string
2711          */
2712         public function uc( $str, $first = false ) {
2713                 if ( $first ) {
2714                         if ( $this->isMultibyte( $str ) ) {
2715                                 return mb_strtoupper( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2716                         } else {
2717                                 return ucfirst( $str );
2718                         }
2719                 } else {
2720                         return $this->isMultibyte( $str ) ? mb_strtoupper( $str ) : strtoupper( $str );
2721                 }
2722         }
2723
2724         /**
2725          * @param string $str
2726          * @return mixed|string
2727          */
2728         function lcfirst( $str ) {
2729                 $o = ord( $str );
2730                 if ( !$o ) {
2731                         return strval( $str );
2732                 } elseif ( $o >= 128 ) {
2733                         return $this->lc( $str, true );
2734                 } elseif ( $o > 96 ) {
2735                         return $str;
2736                 } else {
2737                         $str[0] = strtolower( $str[0] );
2738                         return $str;
2739                 }
2740         }
2741
2742         /**
2743          * @param string $str
2744          * @param bool $first
2745          * @return mixed|string
2746          */
2747         function lc( $str, $first = false ) {
2748                 if ( $first ) {
2749                         if ( $this->isMultibyte( $str ) ) {
2750                                 return mb_strtolower( mb_substr( $str, 0, 1 ) ) . mb_substr( $str, 1 );
2751                         } else {
2752                                 return strtolower( substr( $str, 0, 1 ) ) . substr( $str, 1 );
2753                         }
2754                 } else {
2755                         return $this->isMultibyte( $str ) ? mb_strtolower( $str ) : strtolower( $str );
2756                 }
2757         }
2758
2759         /**
2760          * @param string $str
2761          * @return bool
2762          */
2763         function isMultibyte( $str ) {
2764                 return strlen( $str ) !== mb_strlen( $str );
2765         }
2766
2767         /**
2768          * @param string $str
2769          * @return mixed|string
2770          */
2771         function ucwords( $str ) {
2772                 if ( $this->isMultibyte( $str ) ) {
2773                         $str = $this->lc( $str );
2774
2775                         // regexp to find first letter in each word (i.e. after each space)
2776                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)| ([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2777
2778                         // function to use to capitalize a single char
2779                         return preg_replace_callback(
2780                                 $replaceRegexp,
2781                                 [ $this, 'ucwordsCallbackMB' ],
2782                                 $str
2783                         );
2784                 } else {
2785                         return ucwords( strtolower( $str ) );
2786                 }
2787         }
2788
2789         /**
2790          * capitalize words at word breaks
2791          *
2792          * @param string $str
2793          * @return mixed
2794          */
2795         function ucwordbreaks( $str ) {
2796                 if ( $this->isMultibyte( $str ) ) {
2797                         $str = $this->lc( $str );
2798
2799                         // since \b doesn't work for UTF-8, we explicitely define word break chars
2800                         $breaks = "[ \-\(\)\}\{\.,\?!]";
2801
2802                         // find first letter after word break
2803                         $replaceRegexp = "/^([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)|" .
2804                                 "$breaks([a-z]|[\\xc0-\\xff][\\x80-\\xbf]*)/";
2805
2806                         return preg_replace_callback(
2807                                 $replaceRegexp,
2808                                 [ $this, 'ucwordbreaksCallbackMB' ],
2809                                 $str
2810                         );
2811                 } else {
2812                         return preg_replace_callback(
2813                                 '/\b([\w\x80-\xff]+)\b/',
2814                                 [ $this, 'ucwordbreaksCallbackAscii' ],
2815                                 $str
2816                         );
2817                 }
2818         }
2819
2820         /**
2821          * Return a case-folded representation of $s
2822          *
2823          * This is a representation such that caseFold($s1)==caseFold($s2) if $s1
2824          * and $s2 are the same except for the case of their characters. It is not
2825          * necessary for the value returned to make sense when displayed.
2826          *
2827          * Do *not* perform any other normalisation in this function. If a caller
2828          * uses this function when it should be using a more general normalisation
2829          * function, then fix the caller.
2830          *
2831          * @param string $s
2832          *
2833          * @return string
2834          */
2835         function caseFold( $s ) {
2836                 return $this->uc( $s );
2837         }
2838
2839         /**
2840          * @param string $s
2841          * @return string
2842          */
2843         function checkTitleEncoding( $s ) {
2844                 Assert::parameterType( 'string', $s, '$s' );
2845                 if ( StringUtils::isUtf8( $s ) ) {
2846                         return $s;
2847                 }
2848
2849                 return $this->iconv( $this->fallback8bitEncoding(), 'utf-8', $s );
2850         }
2851
2852         /**
2853          * @return array
2854          */
2855         function fallback8bitEncoding() {
2856                 return self::$dataCache->getItem( $this->mCode, 'fallback8bitEncoding' );
2857         }
2858
2859         /**
2860          * Most writing systems use whitespace to break up words.
2861          * Some languages such as Chinese don't conventionally do this,
2862          * which requires special handling when breaking up words for
2863          * searching etc.
2864          *
2865          * @return bool
2866          */
2867         function hasWordBreaks() {
2868                 return true;
2869         }
2870
2871         /**
2872          * Some languages such as Chinese require word segmentation,
2873          * Specify such segmentation when overridden in derived class.
2874          *
2875          * @param string $string
2876          * @return string
2877          */
2878         function segmentByWord( $string ) {
2879                 return $string;
2880         }
2881
2882         /**
2883          * Some languages have special punctuation need to be normalized.
2884          * Make such changes here.
2885          *
2886          * @param string $string
2887          * @return string
2888          */
2889         function normalizeForSearch( $string ) {
2890                 return self::convertDoubleWidth( $string );
2891         }
2892
2893         /**
2894          * convert double-width roman characters to single-width.
2895          * range: ff00-ff5f ~= 0020-007f
2896          *
2897          * @param string $string
2898          *
2899          * @return string
2900          */
2901         protected static function convertDoubleWidth( $string ) {
2902                 static $full = null;
2903                 static $half = null;
2904
2905                 if ( $full === null ) {
2906                         $fullWidth = "０１２３４５６７８９ＡＢＣＤＥＦＧＨＩＪＫＬＭＮＯＰＱＲＳＴＵＶＷＸＹＺａｂｃｄｅｆｇｈｉｊｋｌｍｎｏｐｑｒｓｔｕｖｗｘｙｚ";
2907                         $halfWidth = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz";
2908                         $full = str_split( $fullWidth, 3 );
2909                         $half = str_split( $halfWidth );
2910                 }
2911
2912                 $string = str_replace( $full, $half, $string );
2913                 return $string;
2914         }
2915
2916         /**
2917          * @param string $string
2918          * @param string $pattern
2919          * @return string
2920          */
2921         protected static function insertSpace( $string, $pattern ) {
2922                 $string = preg_replace( $pattern, " $1 ", $string );
2923                 $string = preg_replace( '/ +/', ' ', $string );
2924                 return $string;
2925         }
2926
2927         /**
2928          * @param array $termsArray
2929          * @return array
2930          */
2931         function convertForSearchResult( $termsArray ) {
2932                 # some languages, e.g. Chinese, need to do a conversion
2933                 # in order for search results to be displayed correctly
2934                 return $termsArray;
2935         }
2936
2937         /**
2938          * Get the first character of a string.
2939          *
2940          * @param string $s
2941          * @return string
2942          */
2943         function firstChar( $s ) {
2944                 $matches = [];
2945                 preg_match(
2946                         '/^([\x00-\x7f]|[\xc0-\xdf][\x80-\xbf]|' .
2947                                 '[\xe0-\xef][\x80-\xbf]{2}|[\xf0-\xf7][\x80-\xbf]{3})/',
2948                         $s,
2949                         $matches
2950                 );
2951
2952                 if ( isset( $matches[1] ) ) {
2953                         if ( strlen( $matches[1] ) != 3 ) {
2954                                 return $matches[1];
2955                         }
2956
2957                         // Break down Hangul syllables to grab the first jamo
2958                         $code = UtfNormal\Utils::utf8ToCodepoint( $matches[1] );
2959                         if ( $code < 0xac00 || 0xd7a4 <= $code ) {
2960                                 return $matches[1];
2961                         } elseif ( $code < 0xb098 ) {
2962                                 return "\u{3131}";
2963                         } elseif ( $code < 0xb2e4 ) {
2964                                 return "\u{3134}";
2965                         } elseif ( $code < 0xb77c ) {
2966                                 return "\u{3137}";
2967                         } elseif ( $code < 0xb9c8 ) {
2968                                 return "\u{3139}";
2969                         } elseif ( $code < 0xbc14 ) {
2970                                 return "\u{3141}";
2971                         } elseif ( $code < 0xc0ac ) {
2972                                 return "\u{3142}";
2973                         } elseif ( $code < 0xc544 ) {
2974                                 return "\u{3145}";
2975                         } elseif ( $code < 0xc790 ) {
2976                                 return "\u{3147}";
2977                         } elseif ( $code < 0xcc28 ) {
2978                                 return "\u{3148}";
2979                         } elseif ( $code < 0xce74 ) {
2980                                 return "\u{314A}";
2981                         } elseif ( $code < 0xd0c0 ) {
2982                                 return "\u{314B}";
2983                         } elseif ( $code < 0xd30c ) {
2984                                 return "\u{314C}";
2985                         } elseif ( $code < 0xd558 ) {
2986                                 return "\u{314D}";
2987                         } else {
2988                                 return "\u{314E}";
2989                         }
2990                 } else {
2991                         return '';
2992                 }
2993         }
2994
2995         /**
2996          * @deprecated No-op since 1.28
2997          */
2998         function initEncoding() {
2999                 wfDeprecated( __METHOD__, '1.28' );
3000                 // No-op.
3001         }
3002
3003         /**
3004          * @param string $s
3005          * @return string
3006          * @deprecated No-op since 1.28
3007          */
3008         function recodeForEdit( $s ) {
3009                 wfDeprecated( __METHOD__, '1.28' );
3010                 return $s;
3011         }
3012
3013         /**
3014          * @param string $s
3015          * @return string
3016          * @deprecated No-op since 1.28
3017          */
3018         function recodeInput( $s ) {
3019                 wfDeprecated( __METHOD__, '1.28' );
3020                 return $s;
3021         }
3022
3023         /**
3024          * Convert a UTF-8 string to normal form C. In Malayalam and Arabic, this
3025          * also cleans up certain backwards-compatible sequences, converting them
3026          * to the modern Unicode equivalent.
3027          *
3028          * This is language-specific for performance reasons only.
3029          *
3030          * @param string $s
3031          *
3032          * @return string
3033          */
3034         public function normalize( $s ) {
3035                 global $wgAllUnicodeFixes;
3036                 $s = UtfNormal\Validator::cleanUp( $s );
3037                 if ( $wgAllUnicodeFixes ) {
3038                         $s = $this->transformUsingPairFile( 'normalize-ar.php', $s );
3039                         $s = $this->transformUsingPairFile( 'normalize-ml.php', $s );
3040                 }
3041
3042                 return $s;
3043         }
3044
3045         /**
3046          * Transform a string using serialized data stored in the given file (which
3047          * must be in the serialized subdirectory of $IP). The file contains pairs
3048          * mapping source characters to destination characters.
3049          *
3050          * The data is cached in process memory. This will go faster if you have the
3051          * FastStringSearch extension.
3052          *
3053          * @param string $file
3054          * @param string $string
3055          *
3056          * @throws MWException
3057          * @return string
3058          */
3059         protected function transformUsingPairFile( $file, $string ) {
3060                 if ( !isset( $this->transformData[$file] ) ) {
3061                         global $IP;
3062                         $data = require "$IP/languages/data/{$file}";
3063                         $this->transformData[$file] = new ReplacementArray( $data );
3064                 }
3065                 return $this->transformData[$file]->replace( $string );
3066         }
3067
3068         /**
3069          * For right-to-left language support
3070          *
3071          * @return bool
3072          */
3073         function isRTL() {
3074                 return self::$dataCache->getItem( $this->mCode, 'rtl' );
3075         }
3076
3077         /**
3078          * Return the correct HTML 'dir' attribute value for this language.
3079          * @return string
3080          */
3081         function getDir() {
3082                 return $this->isRTL() ? 'rtl' : 'ltr';
3083         }
3084
3085         /**
3086          * Return 'left' or 'right' as appropriate alignment for line-start
3087          * for this language's text direction.
3088          *
3089          * Should be equivalent to CSS3 'start' text-align value....
3090          *
3091          * @return string
3092          */
3093         function alignStart() {
3094                 return $this->isRTL() ? 'right' : 'left';
3095         }
3096
3097         /**
3098          * Return 'right' or 'left' as appropriate alignment for line-end
3099          * for this language's text direction.
3100          *
3101          * Should be equivalent to CSS3 'end' text-align value....
3102          *
3103          * @return string
3104          */
3105         function alignEnd() {
3106                 return $this->isRTL() ? 'left' : 'right';
3107         }
3108
3109         /**
3110          * A hidden direction mark (LRM or RLM), depending on the language direction.
3111          * Unlike getDirMark(), this function returns the character as an HTML entity.
3112          * This function should be used when the output is guaranteed to be HTML,
3113          * because it makes the output HTML source code more readable. When
3114          * the output is plain text or can be escaped, getDirMark() should be used.
3115          *
3116          * @param bool $opposite Get the direction mark opposite to your language
3117          * @return string
3118          * @since 1.20
3119          */
3120         function getDirMarkEntity( $opposite = false ) {
3121                 if ( $opposite ) {
3122                         return $this->isRTL() ? '&lrm;' : '&rlm;';
3123                 }
3124                 return $this->isRTL() ? '&rlm;' : '&lrm;';
3125         }
3126
3127         /**
3128          * A hidden direction mark (LRM or RLM), depending on the language direction.
3129          * This function produces them as invisible Unicode characters and
3130          * the output may be hard to read and debug, so it should only be used
3131          * when the output is plain text or can be escaped. When the output is
3132          * HTML, use getDirMarkEntity() instead.
3133          *
3134          * @param bool $opposite Get the direction mark opposite to your language
3135          * @return string
3136          */
3137         function getDirMark( $opposite = false ) {
3138                 $lrm = "\u{200E}"; # LEFT-TO-RIGHT MARK, commonly abbreviated LRM
3139                 $rlm = "\u{200F}"; # RIGHT-TO-LEFT MARK, commonly abbreviated RLM
3140                 if ( $opposite ) {
3141                         return $this->isRTL() ? $lrm : $rlm;
3142                 }
3143                 return $this->isRTL() ? $rlm : $lrm;
3144         }
3145
3146         /**
3147          * @return array
3148          */
3149         function capitalizeAllNouns() {
3150                 return self::$dataCache->getItem( $this->mCode, 'capitalizeAllNouns' );
3151         }
3152
3153         /**
3154          * An arrow, depending on the language direction.
3155          *
3156          * @param string $direction The direction of the arrow: forwards (default),
3157          *   backwards, left, right, up, down.
3158          * @return string
3159          */
3160         function getArrow( $direction = 'forwards' ) {
3161                 switch ( $direction ) {
3162                         case 'forwards':
3163                                 return $this->isRTL() ? '←' : '→';
3164                         case 'backwards':
3165                                 return $this->isRTL() ? '→' : '←';
3166                         case 'left':
3167                                 return '←';
3168                         case 'right':
3169                                 return '→';
3170                         case 'up':
3171                                 return '↑';
3172                         case 'down':
3173                                 return '↓';
3174                 }
3175         }
3176
3177         /**
3178          * To allow "foo[[bar]]" to extend the link over the whole word "foobar"
3179          *
3180          * @return bool
3181          */
3182         function linkPrefixExtension() {
3183                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixExtension' );
3184         }
3185
3186         /**
3187          * Get all magic words from cache.
3188          * @return array
3189          */
3190         function getMagicWords() {
3191                 return self::$dataCache->getItem( $this->mCode, 'magicWords' );
3192         }
3193
3194         /**
3195          * Fill a MagicWord object with data from here
3196          *
3197          * @param MagicWord $mw
3198          */
3199         function getMagic( $mw ) {
3200                 $rawEntry = $this->mMagicExtensions[$mw->mId] ??
3201                         self::$dataCache->getSubitem( $this->mCode, 'magicWords', $mw->mId );
3202
3203                 if ( !is_array( $rawEntry ) ) {
3204                         wfWarn( "\"$rawEntry\" is not a valid magic word for \"$mw->mId\"" );
3205                 } else {
3206                         $mw->mCaseSensitive = $rawEntry[0];
3207                         $mw->mSynonyms = array_slice( $rawEntry, 1 );
3208                 }
3209         }
3210
3211         /**
3212          * Add magic words to the extension array
3213          *
3214          * @param array $newWords
3215          */
3216         function addMagicWordsByLang( $newWords ) {
3217                 $fallbackChain = $this->getFallbackLanguages();
3218                 $fallbackChain = array_reverse( $fallbackChain );
3219                 foreach ( $fallbackChain as $code ) {
3220                         if ( isset( $newWords[$code] ) ) {
3221                                 $this->mMagicExtensions = $newWords[$code] + $this->mMagicExtensions;
3222                         }
3223                 }
3224         }
3225
3226         /**
3227          * Get special page names, as an associative array
3228          *   canonical name => array of valid names, including aliases
3229          * @return array
3230          */
3231         function getSpecialPageAliases() {
3232                 // Cache aliases because it may be slow to load them
3233                 if ( is_null( $this->mExtendedSpecialPageAliases ) ) {
3234                         // Initialise array
3235                         $this->mExtendedSpecialPageAliases =
3236                                 self::$dataCache->getItem( $this->mCode, 'specialPageAliases' );
3237                 }
3238
3239                 return $this->mExtendedSpecialPageAliases;
3240         }
3241
3242         /**
3243          * Italic is unsuitable for some languages
3244          *
3245          * @param string $text The text to be emphasized.
3246          * @return string
3247          */
3248         function emphasize( $text ) {
3249                 return "<em>$text</em>";
3250         }
3251
3252         /**
3253          * Normally we output all numbers in plain en_US style, that is
3254          * 293,291.235 for twohundredninetythreethousand-twohundredninetyone
3255          * point twohundredthirtyfive. However this is not suitable for all
3256          * languages, some such as Bengali (bn) want ২,৯৩,২৯১.২৩৫ and others such as
3257          * Icelandic just want to use commas instead of dots, and dots instead
3258          * of commas like "293.291,235".
3259          *
3260          * An example of this function being called:
3261          * <code>
3262          * wfMessage( 'message' )->numParams( $num )->text()
3263          * </code>
3264          *
3265          * See $separatorTransformTable on MessageIs.php for
3266          * the , => . and . => , implementation.
3267          *
3268          * @todo check if it's viable to use localeconv() for the decimal separator thing.
3269          * @param int|float $number The string to be formatted, should be an integer
3270          *   or a floating point number.
3271          * @param bool $nocommafy Set to true for special numbers like dates
3272          * @return string
3273          */
3274         public function formatNum( $number, $nocommafy = false ) {
3275                 global $wgTranslateNumerals;
3276                 if ( !$nocommafy ) {
3277                         $number = $this->commafy( $number );
3278                         $s = $this->separatorTransformTable();
3279                         if ( $s ) {
3280                                 $number = strtr( $number, $s );
3281                         }
3282                 }
3283
3284                 if ( $wgTranslateNumerals ) {
3285                         $s = $this->digitTransformTable();
3286                         if ( $s ) {
3287                                 $number = strtr( $number, $s );
3288                         }
3289                 }
3290
3291                 return (string)$number;
3292         }
3293
3294         /**
3295          * Front-end for non-commafied formatNum
3296          *
3297          * @param int|float $number The string to be formatted, should be an integer
3298          *        or a floating point number.
3299          * @since 1.21
3300          * @return string
3301          */
3302         public function formatNumNoSeparators( $number ) {
3303                 return $this->formatNum( $number, true );
3304         }
3305
3306         /**
3307          * @param string $number
3308          * @return string
3309          */
3310         public function parseFormattedNumber( $number ) {
3311                 $s = $this->digitTransformTable();
3312                 if ( $s ) {
3313                         // eliminate empty array values such as ''. (T66347)
3314                         $s = array_filter( $s );
3315                         $number = strtr( $number, array_flip( $s ) );
3316                 }
3317
3318                 $s = $this->separatorTransformTable();
3319                 if ( $s ) {
3320                         // eliminate empty array values such as ''. (T66347)
3321                         $s = array_filter( $s );
3322                         $number = strtr( $number, array_flip( $s ) );
3323                 }
3324
3325                 $number = strtr( $number, [ ',' => '' ] );
3326                 return $number;
3327         }
3328
3329         /**
3330          * Adds commas to a given number
3331          * @since 1.19
3332          * @param mixed $number
3333          * @return string
3334          */
3335         function commafy( $number ) {
3336                 $digitGroupingPattern = $this->digitGroupingPattern();
3337                 $minimumGroupingDigits = $this->minimumGroupingDigits();
3338                 if ( $number === null ) {
3339                         return '';
3340                 }
3341
3342                 if ( !$digitGroupingPattern || $digitGroupingPattern === "###,###,###" ) {
3343                         // Default grouping is at thousands, use the same for ###,###,### pattern too.
3344                         // In some languages it's conventional not to insert a thousands separator
3345                         // in numbers that are four digits long (1000-9999).
3346                         if ( $minimumGroupingDigits ) {
3347                                 // Number of '#' characters after last comma in the grouping pattern.
3348                                 // The pattern is hardcoded here, but this would vary for different patterns.
3349                                 $primaryGroupingSize = 3;
3350                                 // Maximum length of a number to suppress digit grouping for.
3351                                 $maximumLength = $minimumGroupingDigits + $primaryGroupingSize - 1;
3352                                 if ( preg_match( '/^\-?\d{1,' . $maximumLength . '}(\.\d+)?$/', $number ) ) {
3353                                         return $number;
3354                                 }
3355                         }
3356                         return strrev( (string)preg_replace( '/(\d{3})(?=\d)(?!\d*\.)/', '$1,', strrev( $number ) ) );
3357                 } else {
3358                         // Ref: http://cldr.unicode.org/translation/number-patterns
3359                         $sign = "";
3360                         if ( intval( $number ) < 0 ) {
3361                                 // For negative numbers apply the algorithm like positive number and add sign.
3362                                 $sign = "-";
3363                                 $number = substr( $number, 1 );
3364                         }
3365                         $integerPart = [];
3366                         $decimalPart = [];
3367                         $numMatches = preg_match_all( "/(#+)/", $digitGroupingPattern, $matches );
3368                         preg_match( "/\d+/", $number, $integerPart );
3369                         preg_match( "/\.\d*/", $number, $decimalPart );
3370                         $groupedNumber = ( count( $decimalPart ) > 0 ) ? $decimalPart[0] : "";
3371                         if ( $groupedNumber === $number ) {
3372                                 // the string does not have any number part. Eg: .12345
3373                                 return $sign . $groupedNumber;
3374                         }
3375                         $start = $end = ( $integerPart ) ? strlen( $integerPart[0] ) : 0;
3376                         while ( $start > 0 ) {
3377                                 $match = $matches[0][$numMatches - 1];
3378                                 $matchLen = strlen( $match );
3379                                 $start = $end - $matchLen;
3380                                 if ( $start < 0 ) {
3381                                         $start = 0;
3382                                 }
3383                                 $groupedNumber = substr( $number, $start, $end - $start ) . $groupedNumber;
3384                                 $end = $start;
3385                                 if ( $numMatches > 1 ) {
3386                                         // use the last pattern for the rest of the number
3387                                         $numMatches--;
3388                                 }
3389                                 if ( $start > 0 ) {
3390                                         $groupedNumber = "," . $groupedNumber;
3391                                 }
3392                         }
3393                         return $sign . $groupedNumber;
3394                 }
3395         }
3396
3397         /**
3398          * @return string
3399          */
3400         function digitGroupingPattern() {
3401                 return self::$dataCache->getItem( $this->mCode, 'digitGroupingPattern' );
3402         }
3403
3404         /**
3405          * @return array
3406          */
3407         function digitTransformTable() {
3408                 return self::$dataCache->getItem( $this->mCode, 'digitTransformTable' );
3409         }
3410
3411         /**
3412          * @return array
3413          */
3414         function separatorTransformTable() {
3415                 return self::$dataCache->getItem( $this->mCode, 'separatorTransformTable' );
3416         }
3417
3418         /**
3419          * @return int|null
3420          */
3421         function minimumGroupingDigits() {
3422                 return self::$dataCache->getItem( $this->mCode, 'minimumGroupingDigits' );
3423         }
3424
3425         /**
3426          * Take a list of strings and build a locale-friendly comma-separated
3427          * list, using the local comma-separator message.
3428          * The last two strings are chained with an "and".
3429          *
3430          * @param string[] $list
3431          * @return string
3432          */
3433         public function listToText( array $list ) {
3434                 $itemCount = count( $list );
3435                 if ( $itemCount < 1 ) {
3436                         return '';
3437                 }
3438                 $text = array_pop( $list );
3439                 if ( $itemCount > 1 ) {
3440                         $and = $this->msg( 'and' )->escaped();
3441                         $space = $this->msg( 'word-separator' )->escaped();
3442                         $comma = '';
3443                         if ( $itemCount > 2 ) {
3444                                 $comma = $this->msg( 'comma-separator' )->escaped();
3445                         }
3446                         $text = implode( $comma, $list ) . $and . $space . $text;
3447                 }
3448                 return $text;
3449         }
3450
3451         /**
3452          * Take a list of strings and build a locale-friendly comma-separated
3453          * list, using the local comma-separator message.
3454          * @param string[] $list Array of strings to put in a comma list
3455          * @return string
3456          */
3457         function commaList( array $list ) {
3458                 return implode(
3459                         wfMessage( 'comma-separator' )->inLanguage( $this )->escaped(),
3460                         $list
3461                 );
3462         }
3463
3464         /**
3465          * Take a list of strings and build a locale-friendly semicolon-separated
3466          * list, using the local semicolon-separator message.
3467          * @param string[] $list Array of strings to put in a semicolon list
3468          * @return string
3469          */
3470         function semicolonList( array $list ) {
3471                 return implode(
3472                         wfMessage( 'semicolon-separator' )->inLanguage( $this )->escaped(),
3473                         $list
3474                 );
3475         }
3476
3477         /**
3478          * Same as commaList, but separate it with the pipe instead.
3479          * @param string[] $list Array of strings to put in a pipe list
3480          * @return string
3481          */
3482         function pipeList( array $list ) {
3483                 return implode(
3484                         wfMessage( 'pipe-separator' )->inLanguage( $this )->escaped(),
3485                         $list
3486                 );
3487         }
3488
3489         /**
3490          * Truncate a string to a specified length in bytes, appending an optional
3491          * string (e.g. for ellipsis)
3492          *
3493          * If $length is negative, the string will be truncated from the beginning
3494          *
3495          * @since 1.31
3496          *
3497          * @param string $string String to truncate
3498          * @param int $length Maximum length in bytes
3499          * @param string $ellipsis String to append to the end of truncated text
3500          * @param bool $adjustLength Subtract length of ellipsis from $length
3501          *
3502          * @return string
3503          */
3504         function truncateForDatabase( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3505                 return $this->truncateInternal(
3506                         $string, $length, $ellipsis, $adjustLength, 'strlen', 'substr'
3507                 );
3508         }
3509
3510         /**
3511          * Truncate a string to a specified number of characters, appending an optional
3512          * string (e.g. for ellipsis).
3513          *
3514          * This provides multibyte version of truncate() method of this class, suitable for truncation
3515          * based on number of characters, instead of number of bytes.
3516          *
3517          * If $length is negative, the string will be truncated from the beginning.
3518          *
3519          * @since 1.31
3520          *
3521          * @param string $string String to truncate
3522          * @param int $length Maximum number of characters
3523          * @param string $ellipsis String to append to the end of truncated text
3524          * @param bool $adjustLength Subtract length of ellipsis from $length
3525          *
3526          * @return string
3527          */
3528         function truncateForVisual( $string, $length, $ellipsis = '...', $adjustLength = true ) {
3529                 // Passing encoding to mb_strlen and mb_substr is optional.
3530                 // Encoding defaults to mb_internal_encoding(), which is set to UTF-8 in Setup.php, so
3531                 // explicit specification of encoding is skipped.
3532                 // Note: Both multibyte methods are callables invoked in truncateInternal.
3533                 return $this->truncateInternal(
3534                         $string, $length, $ellipsis, $adjustLength, 'mb_strlen', 'mb_substr'
3535                 );
3536         }
3537
3538         /**
3539          * Internal method used for truncation. This method abstracts text truncation into
3540          * one common method, allowing users to provide length measurement function and
3541          * function for finding substring.
3542          *
3543          * For usages, see truncateForDatabase and truncateForVisual.
3544          *
3545          * @param string $string String to truncate
3546          * @param int $length Maximum length of final text
3547          * @param string $ellipsis String to append to the end of truncated text
3548          * @param bool $adjustLength Subtract length of ellipsis from $length
3549          * @param callable $measureLength Callable function used for determining the length of text
3550          * @param callable $getSubstring Callable function used for getting the substrings
3551          *
3552          * @return string
3553          */
3554         private function truncateInternal(
3555                 $string, $length, $ellipsis, $adjustLength, $measureLength, $getSubstring
3556         ) {
3557                 if ( !is_callable( $measureLength ) || !is_callable( $getSubstring ) ) {
3558                         throw new InvalidArgumentException( 'Invalid callback provided' );
3559                 }
3560
3561                 # Check if there is no need to truncate
3562                 if ( $measureLength( $string ) <= abs( $length ) ) {
3563                         return $string; // no need to truncate
3564                 }
3565
3566                 # Use the localized ellipsis character
3567                 if ( $ellipsis == '...' ) {
3568                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3569                 }
3570                 if ( $length == 0 ) {
3571                         return $ellipsis; // convention
3572                 }
3573
3574                 $stringOriginal = $string;
3575                 # If ellipsis length is >= $length then we can't apply $adjustLength
3576                 if ( $adjustLength && $measureLength( $ellipsis ) >= abs( $length ) ) {
3577                         $string = $ellipsis; // this can be slightly unexpected
3578                 # Otherwise, truncate and add ellipsis...
3579                 } else {
3580                         $ellipsisLength = $adjustLength ? $measureLength( $ellipsis ) : 0;
3581                         if ( $length > 0 ) {
3582                                 $length -= $ellipsisLength;
3583                                 $string = $getSubstring( $string, 0, $length ); // xyz...
3584                                 $string = $this->removeBadCharLast( $string );
3585                                 $string = rtrim( $string ) . $ellipsis;
3586                         } else {
3587                                 $length += $ellipsisLength;
3588                                 $string = $getSubstring( $string, $length ); // ...xyz
3589                                 $string = $this->removeBadCharFirst( $string );
3590                                 $string = $ellipsis . ltrim( $string );
3591                         }
3592                 }
3593
3594                 # Do not truncate if the ellipsis makes the string longer/equal (T24181).
3595                 # This check is *not* redundant if $adjustLength, due to the single case where
3596                 # LEN($ellipsis) > ABS($limit arg); $stringOriginal could be shorter than $string.
3597                 if ( $measureLength( $string ) < $measureLength( $stringOriginal ) ) {
3598                         return $string;
3599                 } else {
3600                         return $stringOriginal;
3601                 }
3602         }
3603
3604         /**
3605          * Remove bytes that represent an incomplete Unicode character
3606          * at the end of string (e.g. bytes of the char are missing)
3607          *
3608          * @param string $string
3609          * @return string
3610          */
3611         protected function removeBadCharLast( $string ) {
3612                 if ( $string != '' ) {
3613                         $char = ord( $string[strlen( $string ) - 1] );
3614                         $m = [];
3615                         if ( $char >= 0xc0 ) {
3616                                 # We got the first byte only of a multibyte char; remove it.
3617                                 $string = substr( $string, 0, -1 );
3618                         } elseif ( $char >= 0x80 &&
3619                                 // Use the /s modifier (PCRE_DOTALL) so (.*) also matches newlines
3620                                 preg_match( '/^(.*)(?:[\xe0-\xef][\x80-\xbf]|' .
3621                                         '[\xf0-\xf7][\x80-\xbf]{1,2})$/s', $string, $m )
3622                         ) {
3623                                 # We chopped in the middle of a character; remove it
3624                                 $string = $m[1];
3625                         }
3626                 }
3627                 return $string;
3628         }
3629
3630         /**
3631          * Remove bytes that represent an incomplete Unicode character
3632          * at the start of string (e.g. bytes of the char are missing)
3633          *
3634          * @param string $string
3635          * @return string
3636          */
3637         protected function removeBadCharFirst( $string ) {
3638                 if ( $string != '' ) {
3639                         $char = ord( $string[0] );
3640                         if ( $char >= 0x80 && $char < 0xc0 ) {
3641                                 # We chopped in the middle of a character; remove the whole thing
3642                                 $string = preg_replace( '/^[\x80-\xbf]+/', '', $string );
3643                         }
3644                 }
3645                 return $string;
3646         }
3647
3648         /**
3649          * Truncate a string of valid HTML to a specified length in bytes,
3650          * appending an optional string (e.g. for ellipses), and return valid HTML
3651          *
3652          * This is only intended for styled/linked text, such as HTML with
3653          * tags like <span> and <a>, were the tags are self-contained (valid HTML).
3654          * Also, this will not detect things like "display:none" CSS.
3655          *
3656          * Note: since 1.18 you do not need to leave extra room in $length for ellipses.
3657          *
3658          * @param string $text HTML string to truncate
3659          * @param int $length (zero/positive) Maximum length (including ellipses)
3660          * @param string $ellipsis String to append to the truncated text
3661          * @return string
3662          */
3663         function truncateHtml( $text, $length, $ellipsis = '...' ) {
3664                 # Use the localized ellipsis character
3665                 if ( $ellipsis == '...' ) {
3666                         $ellipsis = wfMessage( 'ellipsis' )->inLanguage( $this )->escaped();
3667                 }
3668                 # Check if there is clearly no need to truncate
3669                 if ( $length <= 0 ) {
3670                         return $ellipsis; // no text shown, nothing to format (convention)
3671                 } elseif ( strlen( $text ) <= $length ) {
3672                         return $text; // string short enough even *with* HTML (short-circuit)
3673                 }
3674
3675                 $dispLen = 0; // innerHTML legth so far
3676                 $testingEllipsis = false; // checking if ellipses will make string longer/equal?
3677                 $tagType = 0; // 0-open, 1-close
3678                 $bracketState = 0; // 1-tag start, 2-tag name, 0-neither
3679                 $entityState = 0; // 0-not entity, 1-entity
3680                 $tag = $ret = ''; // accumulated tag name, accumulated result string
3681                 $openTags = []; // open tag stack
3682                 $maybeState = null; // possible truncation state
3683
3684                 $textLen = strlen( $text );
3685                 $neLength = max( 0, $length - strlen( $ellipsis ) ); // non-ellipsis len if truncated
3686                 for ( $pos = 0; true; ++$pos ) {
3687                         # Consider truncation once the display length has reached the maximim.
3688                         # We check if $dispLen > 0 to grab tags for the $neLength = 0 case.
3689                         # Check that we're not in the middle of a bracket/entity...
3690                         if ( $dispLen && $dispLen >= $neLength && $bracketState == 0 && !$entityState ) {
3691                                 if ( !$testingEllipsis ) {
3692                                         $testingEllipsis = true;
3693                                         # Save where we are; we will truncate here unless there turn out to
3694                                         # be so few remaining characters that truncation is not necessary.
3695                                         if ( !$maybeState ) { // already saved? ($neLength = 0 case)
3696                                                 $maybeState = [ $ret, $openTags ]; // save state
3697                                         }
3698                                 } elseif ( $dispLen > $length && $dispLen > strlen( $ellipsis ) ) {
3699                                         # String in fact does need truncation, the truncation point was OK.
3700                                         list( $ret, $openTags ) = $maybeState; // reload state
3701                                         $ret = $this->removeBadCharLast( $ret ); // multi-byte char fix
3702                                         $ret .= $ellipsis; // add ellipsis
3703                                         break;
3704                                 }
3705                         }
3706                         if ( $pos >= $textLen ) {
3707                                 break; // extra iteration just for above checks
3708                         }
3709
3710                         # Read the next char...
3711                         $ch = $text[$pos];
3712                         $lastCh = $pos ? $text[$pos - 1] : '';
3713                         $ret .= $ch; // add to result string
3714                         if ( $ch == '<' ) {
3715                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags ); // for bad HTML
3716                                 $entityState = 0; // for bad HTML
3717                                 $bracketState = 1; // tag started (checking for backslash)
3718                         } elseif ( $ch == '>' ) {
3719                                 $this->truncate_endBracket( $tag, $tagType, $lastCh, $openTags );
3720                                 $entityState = 0; // for bad HTML
3721                                 $bracketState = 0; // out of brackets
3722                         } elseif ( $bracketState == 1 ) {
3723                                 if ( $ch == '/' ) {
3724                                         $tagType = 1; // close tag (e.g. "</span>")
3725                                 } else {
3726                                         $tagType = 0; // open tag (e.g. "<span>")
3727                                         $tag .= $ch;
3728                                 }
3729                                 $bracketState = 2; // building tag name
3730                         } elseif ( $bracketState == 2 ) {
3731                                 if ( $ch != ' ' ) {
3732                                         $tag .= $ch;
3733                                 } else {
3734                                         // Name found (e.g. "<a href=..."), add on tag attributes...
3735                                         $pos += $this->truncate_skip( $ret, $text, "<>", $pos + 1 );
3736                                 }
3737                         } elseif ( $bracketState == 0 ) {
3738                                 if ( $entityState ) {
3739                                         if ( $ch == ';' ) {
3740                                                 $entityState = 0;
3741                                                 $dispLen++; // entity is one displayed char
3742                                         }
3743                                 } else {
3744                                         if ( $neLength == 0 && !$maybeState ) {
3745                                                 // Save state without $ch. We want to *hit* the first
3746                                                 // display char (to get tags) but not *use* it if truncating.
3747                                                 $maybeState = [ substr( $ret, 0, -1 ), $openTags ];
3748                                         }
3749                                         if ( $ch == '&' ) {
3750                                                 $entityState = 1; // entity found, (e.g. "&#160;")
3751                                         } else {
3752                                                 $dispLen++; // this char is displayed
3753                                                 // Add the next $max display text chars after this in one swoop...
3754                                                 $max = ( $testingEllipsis ? $length : $neLength ) - $dispLen;
3755                                                 $skipped = $this->truncate_skip( $ret, $text, "<>&", $pos + 1, $max );
3756                                                 $dispLen += $skipped;
3757                                                 $pos += $skipped;
3758                                         }
3759                                 }
3760                         }
3761                 }
3762                 // Close the last tag if left unclosed by bad HTML
3763                 $this->truncate_endBracket( $tag, $text[$textLen - 1], $tagType, $openTags );
3764                 while ( count( $openTags ) > 0 ) {
3765                         $ret .= '</' . array_pop( $openTags ) . '>'; // close open tags
3766                 }
3767                 return $ret;
3768         }
3769
3770         /**
3771          * truncateHtml() helper function
3772          * like strcspn() but adds the skipped chars to $ret
3773          *
3774          * @param string $ret
3775          * @param string $text
3776          * @param string $search
3777          * @param int $start
3778          * @param null|int $len
3779          * @return int
3780          */
3781         private function truncate_skip( &$ret, $text, $search, $start, $len = null ) {
3782                 if ( $len === null ) {
3783                         $len = -1; // -1 means "no limit" for strcspn
3784                 } elseif ( $len < 0 ) {
3785                         $len = 0; // sanity
3786                 }
3787                 $skipCount = 0;
3788                 if ( $start < strlen( $text ) ) {
3789                         $skipCount = strcspn( $text, $search, $start, $len );
3790                         $ret .= substr( $text, $start, $skipCount );
3791                 }
3792                 return $skipCount;
3793         }
3794
3795         /**
3796          * truncateHtml() helper function
3797          * (a) push or pop $tag from $openTags as needed
3798          * (b) clear $tag value
3799          * @param string &$tag Current HTML tag name we are looking at
3800          * @param int $tagType (0-open tag, 1-close tag)
3801          * @param string $lastCh Character before the '>' that ended this tag
3802          * @param array &$openTags Open tag stack (not accounting for $tag)
3803          */
3804         private function truncate_endBracket( &$tag, $tagType, $lastCh, &$openTags ) {
3805                 $tag = ltrim( $tag );
3806                 if ( $tag != '' ) {
3807                         if ( $tagType == 0 && $lastCh != '/' ) {
3808                                 $openTags[] = $tag; // tag opened (didn't close itself)
3809                         } elseif ( $tagType == 1 ) {
3810                                 if ( $openTags && $tag == $openTags[count( $openTags ) - 1] ) {
3811                                         array_pop( $openTags ); // tag closed
3812                                 }
3813                         }
3814                         $tag = '';
3815                 }
3816         }
3817
3818         /**
3819          * Grammatical transformations, needed for inflected languages
3820          * Invoked by putting {{grammar:case|word}} in a message
3821          *
3822          * @param string $word
3823          * @param string $case
3824          * @return string
3825          */
3826         function convertGrammar( $word, $case ) {
3827                 global $wgGrammarForms;
3828                 if ( isset( $wgGrammarForms[$this->getCode()][$case][$word] ) ) {
3829                         return $wgGrammarForms[$this->getCode()][$case][$word];
3830                 }
3831
3832                 $grammarTransformations = $this->getGrammarTransformations();
3833
3834                 if ( isset( $grammarTransformations[$case] ) ) {
3835                         $forms = $grammarTransformations[$case];
3836
3837                         // Some names of grammar rules are aliases for other rules.
3838                         // In such cases the value is a string rather than object,
3839                         // so load the actual rules.
3840                         if ( is_string( $forms ) ) {
3841                                 $forms = $grammarTransformations[$forms];
3842                         }
3843
3844                         foreach ( array_values( $forms ) as $rule ) {
3845                                 $form = $rule[0];
3846
3847                                 if ( $form === '@metadata' ) {
3848                                         continue;
3849                                 }
3850
3851                                 $replacement = $rule[1];
3852
3853                                 $regex = '/' . addcslashes( $form, '/' ) . '/u';
3854                                 $patternMatches = preg_match( $regex, $word );
3855
3856                                 if ( $patternMatches === false ) {
3857                                         wfLogWarning(
3858                                                 'An error occurred while processing grammar. ' .
3859                                                 "Word: '$word'. Regex: /$form/."
3860                                         );
3861                                 } elseif ( $patternMatches === 1 ) {
3862                                         $word = preg_replace( $regex, $replacement, $word );
3863
3864                                         break;
3865                                 }
3866                         }
3867                 }
3868
3869                 return $word;
3870         }
3871
3872         /**
3873          * Get the grammar forms for the content language
3874          * @return array Array of grammar forms
3875          * @since 1.20
3876          */
3877         function getGrammarForms() {
3878                 global $wgGrammarForms;
3879                 if ( isset( $wgGrammarForms[$this->getCode()] )
3880                         && is_array( $wgGrammarForms[$this->getCode()] )
3881                 ) {
3882                         return $wgGrammarForms[$this->getCode()];
3883                 }
3884
3885                 return [];
3886         }
3887
3888         /**
3889          * Get the grammar transformations data for the language.
3890          * Used like grammar forms, with {{GRAMMAR}} and cases,
3891          * but uses pairs of regexes and replacements instead of code.
3892          *
3893          * @return array[] Array of grammar transformations.
3894          * @throws MWException
3895          * @since 1.28
3896          */
3897         public function getGrammarTransformations() {
3898                 $languageCode = $this->getCode();
3899
3900                 if ( self::$grammarTransformations === null ) {
3901                         self::$grammarTransformations = new MapCacheLRU( 10 );
3902                 }
3903
3904                 if ( self::$grammarTransformations->has( $languageCode ) ) {
3905                         return self::$grammarTransformations->get( $languageCode );
3906                 }
3907
3908                 $data = [];
3909
3910                 $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
3911                 if ( is_readable( $grammarDataFile ) ) {
3912                         $data = FormatJson::decode(
3913                                 file_get_contents( $grammarDataFile ),
3914                                 true
3915                         );
3916
3917                         if ( $data === null ) {
3918                                 throw new MWException( "Invalid grammar data for \"$languageCode\"." );
3919                         }
3920
3921                         self::$grammarTransformations->set( $languageCode, $data );
3922                 }
3923
3924                 return $data;
3925         }
3926
3927         /**
3928          * Provides an alternative text depending on specified gender.
3929          * Usage {{gender:username|masculine|feminine|unknown}}.
3930          * username is optional, in which case the gender of current user is used,
3931          * but only in (some) interface messages; otherwise default gender is used.
3932          *
3933          * If no forms are given, an empty string is returned. If only one form is
3934          * given, it will be returned unconditionally. These details are implied by
3935          * the caller and cannot be overridden in subclasses.
3936          *
3937          * If three forms are given, the default is to use the third (unknown) form.
3938          * If fewer than three forms are given, the default is to use the first (masculine) form.
3939          * These details can be overridden in subclasses.
3940          *
3941          * @param string $gender
3942          * @param array $forms
3943          *
3944          * @return string
3945          */
3946         function gender( $gender, $forms ) {
3947                 if ( !count( $forms ) ) {
3948                         return '';
3949                 }
3950                 $forms = $this->preConvertPlural( $forms, 2 );
3951                 if ( $gender === 'male' ) {
3952                         return $forms[0];
3953                 }
3954                 if ( $gender === 'female' ) {
3955                         return $forms[1];
3956                 }
3957                 return $forms[2] ?? $forms[0];
3958         }
3959
3960         /**
3961          * Plural form transformations, needed for some languages.
3962          * For example, there are 3 form of plural in Russian and Polish,
3963          * depending on "count mod 10". See [[w:Plural]]
3964          * For English it is pretty simple.
3965          *
3966          * Invoked by putting {{plural:count|wordform1|wordform2}}
3967          * or {{plural:count|wordform1|wordform2|wordform3}}
3968          *
3969          * Example: {{plural:{{NUMBEROFARTICLES}}|article|articles}}
3970          *
3971          * @param int $count Non-localized number
3972          * @param array $forms Different plural forms
3973          * @return string Correct form of plural for $count in this language
3974          */
3975         function convertPlural( $count, $forms ) {
3976                 // Handle explicit n=pluralform cases
3977                 $forms = $this->handleExplicitPluralForms( $count, $forms );
3978                 if ( is_string( $forms ) ) {
3979                         return $forms;
3980                 }
3981                 if ( !count( $forms ) ) {
3982                         return '';
3983                 }
3984
3985                 $pluralForm = $this->getPluralRuleIndexNumber( $count );
3986                 $pluralForm = min( $pluralForm, count( $forms ) - 1 );
3987                 return $forms[$pluralForm];
3988         }
3989
3990         /**
3991          * Handles explicit plural forms for Language::convertPlural()
3992          *
3993          * In {{PLURAL:$1|0=nothing|one|many}}, 0=nothing will be returned if $1 equals zero.
3994          * If an explicitly defined plural form matches the $count, then
3995          * string value returned, otherwise array returned for further consideration
3996          * by CLDR rules or overridden convertPlural().
3997          *
3998          * @since 1.23
3999          *
4000          * @param int $count Non-localized number
4001          * @param array $forms Different plural forms
4002          *
4003          * @return array|string
4004          */
4005         protected function handleExplicitPluralForms( $count, array $forms ) {
4006                 foreach ( $forms as $index => $form ) {
4007                         if ( preg_match( '/\d+=/i', $form ) ) {
4008                                 $pos = strpos( $form, '=' );
4009                                 if ( substr( $form, 0, $pos ) === (string)$count ) {
4010                                         return substr( $form, $pos + 1 );
4011                                 }
4012                                 unset( $forms[$index] );
4013                         }
4014                 }
4015                 return array_values( $forms );
4016         }
4017
4018         /**
4019          * Checks that convertPlural was given an array and pads it to requested
4020          * amount of forms by copying the last one.
4021          *
4022          * @param array $forms Array of forms given to convertPlural
4023          * @param int $count How many forms should there be at least
4024          * @return array Padded array of forms or an exception if not an array
4025          */
4026         protected function preConvertPlural( /* Array */ $forms, $count ) {
4027                 while ( count( $forms ) < $count ) {
4028                         $forms[] = $forms[count( $forms ) - 1];
4029                 }
4030                 return $forms;
4031         }
4032
4033         /**
4034          * Wraps argument with unicode control characters for directionality safety
4035          *
4036          * This solves the problem where directionality-neutral characters at the edge of
4037          * the argument string get interpreted with the wrong directionality from the
4038          * enclosing context, giving renderings that look corrupted like "(Ben_(WMF".
4039          *
4040          * The wrapping is LRE...PDF or RLE...PDF, depending on the detected
4041          * directionality of the argument string, using the BIDI algorithm's own "First
4042          * strong directional codepoint" rule. Essentially, this works round the fact that
4043          * there is no embedding equivalent of U+2068 FSI (isolation with heuristic
4044          * direction inference). The latter is cleaner but still not widely supported.
4045          *
4046          * @param string $text Text to wrap
4047          * @return string Text, wrapped in LRE...PDF or RLE...PDF or nothing
4048          */
4049         public function embedBidi( $text = '' ) {
4050                 $dir = self::strongDirFromContent( $text );
4051                 if ( $dir === 'ltr' ) {
4052                         // Wrap in LEFT-TO-RIGHT EMBEDDING ... POP DIRECTIONAL FORMATTING
4053                         return self::$lre . $text . self::$pdf;
4054                 }
4055                 if ( $dir === 'rtl' ) {
4056                         // Wrap in RIGHT-TO-LEFT EMBEDDING ... POP DIRECTIONAL FORMATTING
4057                         return self::$rle . $text . self::$pdf;
4058                 }
4059                 // No strong directionality: do not wrap
4060                 return $text;
4061         }
4062
4063         /**
4064          * @todo Maybe translate block durations.  Note that this function is somewhat misnamed: it
4065          * deals with translating the *duration* ("1 week", "4 days", etc), not the expiry time
4066          * (which is an absolute timestamp). Please note: do NOT add this blindly, as it is used
4067          * on old expiry lengths recorded in log entries. You'd need to provide the start date to
4068          * match up with it.
4069          *
4070          * @param string $str The validated block duration in English
4071          * @param User|null $user User object to use timezone from or null for $wgUser
4072          * @param int $now Current timestamp, for formatting relative block durations
4073          * @return string Somehow translated block duration
4074          * @see LanguageFi.php for example implementation
4075          */
4076         function translateBlockExpiry( $str, User $user = null, $now = 0 ) {
4077                 $duration = SpecialBlock::getSuggestedDurations( $this );
4078                 foreach ( $duration as $show => $value ) {
4079                         if ( strcmp( $str, $value ) == 0 ) {
4080                                 return htmlspecialchars( trim( $show ) );
4081                         }
4082                 }
4083
4084                 if ( wfIsInfinity( $str ) ) {
4085                         foreach ( $duration as $show => $value ) {
4086                                 if ( wfIsInfinity( $value ) ) {
4087                                         return htmlspecialchars( trim( $show ) );
4088                                 }
4089                         }
4090                 }
4091
4092                 // If all else fails, return a standard duration or timestamp description.
4093                 $time = strtotime( $str, $now );
4094                 if ( $time === false ) { // Unknown format. Return it as-is in case.
4095                         return $str;
4096                 } elseif ( $time !== strtotime( $str, $now + 1 ) ) { // It's a relative timestamp.
4097                         // The result differs based on current time, so the difference
4098                         // is a fixed duration length.
4099                         return $this->formatDuration( $time - $now );
4100                 } else { // It's an absolute timestamp.
4101                         if ( $time === 0 ) {
4102                                 // wfTimestamp() handles 0 as current time instead of epoch.
4103                                 $time = '19700101000000';
4104                         }
4105                         if ( $user ) {
4106                                 return $this->userTimeAndDate( $time, $user );
4107                         }
4108                         return $this->timeanddate( $time );
4109                 }
4110         }
4111
4112         /**
4113          * languages like Chinese need to be segmented in order for the diff
4114          * to be of any use
4115          *
4116          * @param string $text
4117          * @return string
4118          */
4119         public function segmentForDiff( $text ) {
4120                 return $text;
4121         }
4122
4123         /**
4124          * and unsegment to show the result
4125          *
4126          * @param string $text
4127          * @return string
4128          */
4129         public function unsegmentForDiff( $text ) {
4130                 return $text;
4131         }
4132
4133         /**
4134          * Return the LanguageConverter used in the Language
4135          *
4136          * @since 1.19
4137          * @return LanguageConverter
4138          */
4139         public function getConverter() {
4140                 return $this->mConverter;
4141         }
4142
4143         /**
4144          * convert text to a variant
4145          *
4146          * @param string $text text to convert
4147          * @param string|bool $variant variant to convert to, or false to use the user's preferred
4148          *      variant (if logged in), or the project default variant
4149          * @return string the converted string
4150          */
4151         public function autoConvert( $text, $variant = false ) {
4152                 return $this->mConverter->autoConvert( $text, $variant );
4153         }
4154
4155         /**
4156          * convert text to all supported variants
4157          *
4158          * @param string $text
4159          * @return array
4160          */
4161         public function autoConvertToAllVariants( $text ) {
4162                 return $this->mConverter->autoConvertToAllVariants( $text );
4163         }
4164
4165         /**
4166          * convert text to different variants of a language.
4167          *
4168          * @warning Glossary state is maintained between calls. This means
4169          *  if you pass unescaped text to this method it can cause an XSS
4170          *  in later calls to this method, even if the later calls have properly
4171          *  escaped the input. Never feed this method user controlled text that
4172          *  is not properly escaped!
4173          * @param string $text Content that has been already escaped for use in HTML
4174          * @return string HTML
4175          */
4176         public function convert( $text ) {
4177                 return $this->mConverter->convert( $text );
4178         }
4179
4180         /**
4181          * Convert a Title object to a string in the preferred variant
4182          *
4183          * @param Title $title
4184          * @return string
4185          */
4186         public function convertTitle( $title ) {
4187                 return $this->mConverter->convertTitle( $title );
4188         }
4189
4190         /**
4191          * Convert a namespace index to a string in the preferred variant
4192          *
4193          * @param int $ns namespace index (https://www.mediawiki.org/wiki/Manual:Namespace)
4194          * @param string|null $variant variant to convert to, or null to use the user's preferred
4195          *      variant (if logged in), or the project default variant
4196          * @return string a string representation of the namespace
4197          */
4198         public function convertNamespace( $ns, $variant = null ) {
4199                 return $this->mConverter->convertNamespace( $ns, $variant );
4200         }
4201
4202         /**
4203          * Check if this is a language with variants
4204          *
4205          * @return bool
4206          */
4207         public function hasVariants() {
4208                 return count( $this->getVariants() ) > 1;
4209         }
4210
4211         /**
4212          * Strict check if the language has the specific variant.
4213          *
4214          * Compare to LanguageConverter::validateVariant() which does a more
4215          * lenient check and attempts to coerce the given code to a valid one.
4216          *
4217          * @since 1.19
4218          * @param string $variant
4219          * @return bool
4220          */
4221         public function hasVariant( $variant ) {
4222                 return $variant && ( $variant === $this->mConverter->validateVariant( $variant ) );
4223         }
4224
4225         /**
4226          * Perform output conversion on a string, and encode for safe HTML output.
4227          * @param string $text Text to be converted
4228          * @return string
4229          * @todo this should get integrated somewhere sane
4230          */
4231         public function convertHtml( $text ) {
4232                 return htmlspecialchars( $this->convert( $text ) );
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( $title );
4319         }
4320
4321         /**
4322          * A regular expression to match legal word-trailing characters
4323          * which should be merged onto a link of the form [[foo]]bar.
4324          *
4325          * @return string
4326          */
4327         public function linkTrail() {
4328                 return self::$dataCache->getItem( $this->mCode, 'linkTrail' );
4329         }
4330
4331         /**
4332          * A regular expression character set to match legal word-prefixing
4333          * characters which should be merged onto a link of the form foo[[bar]].
4334          *
4335          * @return string
4336          */
4337         public function linkPrefixCharset() {
4338                 return self::$dataCache->getItem( $this->mCode, 'linkPrefixCharset' );
4339         }
4340
4341         /**
4342          * Get the "parent" language which has a converter to convert a "compatible" language
4343          * (in another variant) to this language (eg. zh for zh-cn, but not en for en-gb).
4344          *
4345          * @return Language|null
4346          * @since 1.22
4347          */
4348         public function getParentLanguage() {
4349                 if ( $this->mParentLanguage !== false ) {
4350                         return $this->mParentLanguage;
4351                 }
4352
4353                 $code = explode( '-', $this->getCode() )[0];
4354                 if ( !in_array( $code, LanguageConverter::$languagesWithVariants ) ) {
4355                         $this->mParentLanguage = null;
4356                         return null;
4357                 }
4358                 $lang = self::factory( $code );
4359                 if ( !$lang->hasVariant( $this->getCode() ) ) {
4360                         $this->mParentLanguage = null;
4361                         return null;
4362                 }
4363
4364                 $this->mParentLanguage = $lang;
4365                 return $lang;
4366         }
4367
4368         /**
4369          * Compare with an other language object
4370          *
4371          * @since 1.28
4372          * @param Language $lang
4373          * @return bool
4374          */
4375         public function equals( Language $lang ) {
4376                 return $lang === $this || $lang->getCode() === $this->mCode;
4377         }
4378
4379         /**
4380          * Get the internal language code for this language object
4381          *
4382          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4383          * htmlspecialchars() or similar
4384          *
4385          * @return string
4386          */
4387         public function getCode() {
4388                 return $this->mCode;
4389         }
4390
4391         /**
4392          * Get the code in BCP 47 format which we can use
4393          * inside of html lang="" tags.
4394          *
4395          * NOTE: The return value of this function is NOT HTML-safe and must be escaped with
4396          * htmlspecialchars() or similar.
4397          *
4398          * @since 1.19
4399          * @return string
4400          */
4401         public function getHtmlCode() {
4402                 if ( is_null( $this->mHtmlCode ) ) {
4403                         $this->mHtmlCode = LanguageCode::bcp47( $this->getCode() );
4404                 }
4405                 return $this->mHtmlCode;
4406         }
4407
4408         /**
4409          * @param string $code
4410          * @deprecated since 1.32, use Language::factory to create a new object instead.
4411          */
4412         public function setCode( $code ) {
4413                 wfDeprecated( __METHOD__, '1.32' );
4414                 $this->mCode = $code;
4415                 // Ensure we don't leave incorrect cached data lying around
4416                 $this->mHtmlCode = null;
4417                 $this->mParentLanguage = false;
4418         }
4419
4420         /**
4421          * Get the language code from a file name. Inverse of getFileName()
4422          * @param string $filename $prefix . $languageCode . $suffix
4423          * @param string $prefix Prefix before the language code
4424          * @param string $suffix Suffix after the language code
4425          * @return string Language code, or false if $prefix or $suffix isn't found
4426          */
4427         public static function getCodeFromFileName( $filename, $prefix = 'Language', $suffix = '.php' ) {
4428                 $m = null;
4429                 preg_match( '/' . preg_quote( $prefix, '/' ) . '([A-Z][a-z_]+)' .
4430                         preg_quote( $suffix, '/' ) . '/', $filename, $m );
4431                 if ( !count( $m ) ) {
4432                         return false;
4433                 }
4434                 return str_replace( '_', '-', strtolower( $m[1] ) );
4435         }
4436
4437         /**
4438          * @param string $code
4439          * @param bool $fallback Whether we're going through language fallback chain
4440          * @return string Name of the language class
4441          */
4442         public static function classFromCode( $code, $fallback = true ) {
4443                 if ( $fallback && $code == 'en' ) {
4444                         return 'Language';
4445                 } else {
4446                         return 'Language' . str_replace( '-', '_', ucfirst( $code ) );
4447                 }
4448         }
4449
4450         /**
4451          * Get the name of a file for a certain language code
4452          * @param string $prefix Prepend this to the filename
4453          * @param string $code Language code
4454          * @param string $suffix Append this to the filename
4455          * @throws MWException
4456          * @return string $prefix . $mangledCode . $suffix
4457          */
4458         public static function getFileName( $prefix, $code, $suffix = '.php' ) {
4459                 if ( !self::isValidBuiltInCode( $code ) ) {
4460                         throw new MWException( "Invalid language code \"$code\"" );
4461                 }
4462
4463                 return $prefix . str_replace( '-', '_', ucfirst( $code ) ) . $suffix;
4464         }
4465
4466         /**
4467          * @param string $code
4468          * @return string
4469          */
4470         public static function getMessagesFileName( $code ) {
4471                 global $IP;
4472                 $file = self::getFileName( "$IP/languages/messages/Messages", $code, '.php' );
4473                 Hooks::run( 'Language::getMessagesFileName', [ $code, &$file ] );
4474                 return $file;
4475         }
4476
4477         /**
4478          * @param string $code
4479          * @return string
4480          * @throws MWException
4481          * @since 1.23
4482          */
4483         public static function getJsonMessagesFileName( $code ) {
4484                 global $IP;
4485
4486                 if ( !self::isValidBuiltInCode( $code ) ) {
4487                         throw new MWException( "Invalid language code \"$code\"" );
4488                 }
4489
4490                 return "$IP/languages/i18n/$code.json";
4491         }
4492
4493         /**
4494          * Get the first fallback for a given language.
4495          *
4496          * @param string $code
4497          *
4498          * @return bool|string
4499          */
4500         public static function getFallbackFor( $code ) {
4501                 $fallbacks = self::getFallbacksFor( $code );
4502                 if ( $fallbacks ) {
4503                         return $fallbacks[0];
4504                 }
4505                 return false;
4506         }
4507
4508         /**
4509          * Get the ordered list of fallback languages.
4510          *
4511          * @since 1.19
4512          * @param string $code Language code
4513          * @param int $mode Fallback mode, either MESSAGES_FALLBACKS (which always falls back to 'en'),
4514          * or STRICT_FALLBACKS (whic honly falls back to 'en' when explicitly defined)
4515          * @throws MWException
4516          * @return array List of language codes
4517          */
4518         public static function getFallbacksFor( $code, $mode = self::MESSAGES_FALLBACKS ) {
4519                 if ( $code === 'en' || !self::isValidBuiltInCode( $code ) ) {
4520                         return [];
4521                 }
4522                 switch ( $mode ) {
4523                         case self::MESSAGES_FALLBACKS:
4524                                 // For unknown languages, fallbackSequence returns an empty array,
4525                                 // hardcode fallback to 'en' in that case as English messages are
4526                                 // always defined.
4527                                 return self::getLocalisationCache()->getItem( $code, 'fallbackSequence' ) ?: [ 'en' ];
4528                         case self::STRICT_FALLBACKS:
4529                                 // Use this mode when you don't want to fallback to English unless
4530                                 // explicitly defined, for example when you have language-variant icons
4531                                 // and an international language-independent fallback.
4532                                 return self::getLocalisationCache()->getItem( $code, 'originalFallbackSequence' );
4533                         default:
4534                                 throw new MWException( "Invalid fallback mode \"$mode\"" );
4535                 }
4536         }
4537
4538         /**
4539          * Get the ordered list of fallback languages, ending with the fallback
4540          * language chain for the site language.
4541          *
4542          * @since 1.22
4543          * @param string $code Language code
4544          * @return array Array( fallbacks, site fallbacks )
4545          */
4546         public static function getFallbacksIncludingSiteLanguage( $code ) {
4547                 global $wgLanguageCode;
4548
4549                 // Usually, we will only store a tiny number of fallback chains, so we
4550                 // keep them in static memory.
4551                 $cacheKey = "{$code}-{$wgLanguageCode}";
4552
4553                 if ( !array_key_exists( $cacheKey, self::$fallbackLanguageCache ) ) {
4554                         $fallbacks = self::getFallbacksFor( $code );
4555
4556                         // Append the site's fallback chain, including the site language itself
4557                         $siteFallbacks = self::getFallbacksFor( $wgLanguageCode );
4558                         array_unshift( $siteFallbacks, $wgLanguageCode );
4559
4560                         // Eliminate any languages already included in the chain
4561                         $siteFallbacks = array_diff( $siteFallbacks, $fallbacks );
4562
4563                         self::$fallbackLanguageCache[$cacheKey] = [ $fallbacks, $siteFallbacks ];
4564                 }
4565                 return self::$fallbackLanguageCache[$cacheKey];
4566         }
4567
4568         /**
4569          * Get all messages for a given language
4570          * WARNING: this may take a long time. If you just need all message *keys*
4571          * but need the *contents* of only a few messages, consider using getMessageKeysFor().
4572          *
4573          * @param string $code
4574          *
4575          * @return array
4576          */
4577         public static function getMessagesFor( $code ) {
4578                 return self::getLocalisationCache()->getItem( $code, 'messages' );
4579         }
4580
4581         /**
4582          * Get a message for a given language
4583          *
4584          * @param string $key
4585          * @param string $code
4586          *
4587          * @return string
4588          */
4589         public static function getMessageFor( $key, $code ) {
4590                 return self::getLocalisationCache()->getSubitem( $code, 'messages', $key );
4591         }
4592
4593         /**
4594          * Get all message keys for a given language. This is a faster alternative to
4595          * array_keys( Language::getMessagesFor( $code ) )
4596          *
4597          * @since 1.19
4598          * @param string $code Language code
4599          * @return array Array of message keys (strings)
4600          */
4601         public static function getMessageKeysFor( $code ) {
4602                 return self::getLocalisationCache()->getSubitemList( $code, 'messages' );
4603         }
4604
4605         /**
4606          * @param string $talk
4607          * @return string
4608          */
4609         function fixVariableInNamespace( $talk ) {
4610                 if ( strpos( $talk, '$1' ) === false ) {
4611                         return $talk;
4612                 }
4613
4614                 global $wgMetaNamespace;
4615                 $talk = str_replace( '$1', $wgMetaNamespace, $talk );
4616
4617                 # Allow grammar transformations
4618                 # Allowing full message-style parsing would make simple requests
4619                 # such as action=raw much more expensive than they need to be.
4620                 # This will hopefully cover most cases.
4621                 $talk = preg_replace_callback( '/{{grammar:(.*?)\|(.*?)}}/i',
4622                         [ $this, 'replaceGrammarInNamespace' ], $talk );
4623                 return str_replace( ' ', '_', $talk );
4624         }
4625
4626         /**
4627          * @param string $m
4628          * @return string
4629          */
4630         function replaceGrammarInNamespace( $m ) {
4631                 return $this->convertGrammar( trim( $m[2] ), trim( $m[1] ) );
4632         }
4633
4634         /**
4635          * Decode an expiry (block, protection, etc) which has come from the DB
4636          *
4637          * @param string $expiry Database expiry String
4638          * @param bool|int $format True to process using language functions, or TS_ constant
4639          *     to return the expiry in a given timestamp
4640          * @param string $infinity If $format is not true, use this string for infinite expiry
4641          * @return string
4642          * @since 1.18
4643          */
4644         public function formatExpiry( $expiry, $format = true, $infinity = 'infinity' ) {
4645                 static $dbInfinity;
4646                 if ( $dbInfinity === null ) {
4647                         $dbInfinity = wfGetDB( DB_REPLICA )->getInfinity();
4648                 }
4649
4650                 if ( $expiry == '' || $expiry === 'infinity' || $expiry == $dbInfinity ) {
4651                         return $format === true
4652                                 ? $this->getMessageFromDB( 'infiniteblock' )
4653                                 : $infinity;
4654                 } else {
4655                         return $format === true
4656                                 ? $this->timeanddate( $expiry, /* User preference timezone */ true )
4657                                 : wfTimestamp( $format, $expiry );
4658                 }
4659         }
4660
4661         /**
4662          * Formats a time given in seconds into a string representation of that time.
4663          *
4664          * @param int|float $seconds
4665          * @param array $format An optional argument that formats the returned string in different ways:
4666          *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
4667          *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
4668          *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
4669          *     and friends.
4670          * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
4671          *     or 'avoidminutes'.
4672          * @return string
4673          */
4674         function formatTimePeriod( $seconds, $format = [] ) {
4675                 if ( !is_array( $format ) ) {
4676                         $format = [ 'avoid' => $format ]; // For backwards compatibility
4677                 }
4678                 if ( !isset( $format['avoid'] ) ) {
4679                         $format['avoid'] = false;
4680                 }
4681                 if ( !isset( $format['noabbrevs'] ) ) {
4682                         $format['noabbrevs'] = false;
4683                 }
4684                 $secondsMsg = wfMessage(
4685                         $format['noabbrevs'] ? 'seconds' : 'seconds-abbrev' )->inLanguage( $this );
4686                 $minutesMsg = wfMessage(
4687                         $format['noabbrevs'] ? 'minutes' : 'minutes-abbrev' )->inLanguage( $this );
4688                 $hoursMsg = wfMessage(
4689                         $format['noabbrevs'] ? 'hours' : 'hours-abbrev' )->inLanguage( $this );
4690                 $daysMsg = wfMessage(
4691                         $format['noabbrevs'] ? 'days' : 'days-abbrev' )->inLanguage( $this );
4692
4693                 if ( round( $seconds * 10 ) < 100 ) {
4694                         $s = $this->formatNum( sprintf( "%.1f", round( $seconds * 10 ) / 10 ) );
4695                         $s = $secondsMsg->params( $s )->text();
4696                 } elseif ( round( $seconds ) < 60 ) {
4697                         $s = $this->formatNum( round( $seconds ) );
4698                         $s = $secondsMsg->params( $s )->text();
4699                 } elseif ( round( $seconds ) < 3600 ) {
4700                         $minutes = floor( $seconds / 60 );
4701                         $secondsPart = round( fmod( $seconds, 60 ) );
4702                         if ( $secondsPart == 60 ) {
4703                                 $secondsPart = 0;
4704                                 $minutes++;
4705                         }
4706                         $s = $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4707                         $s .= ' ';
4708                         $s .= $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4709                 } elseif ( round( $seconds ) <= 2 * 86400 ) {
4710                         $hours = floor( $seconds / 3600 );
4711                         $minutes = floor( ( $seconds - $hours * 3600 ) / 60 );
4712                         $secondsPart = round( $seconds - $hours * 3600 - $minutes * 60 );
4713                         if ( $secondsPart == 60 ) {
4714                                 $secondsPart = 0;
4715                                 $minutes++;
4716                         }
4717                         if ( $minutes == 60 ) {
4718                                 $minutes = 0;
4719                                 $hours++;
4720                         }
4721                         $s = $hoursMsg->params( $this->formatNum( $hours ) )->text();
4722                         $s .= ' ';
4723                         $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4724                         if ( !in_array( $format['avoid'], [ 'avoidseconds', 'avoidminutes' ] ) ) {
4725                                 $s .= ' ' . $secondsMsg->params( $this->formatNum( $secondsPart ) )->text();
4726                         }
4727                 } else {
4728                         $days = floor( $seconds / 86400 );
4729                         if ( $format['avoid'] === 'avoidminutes' ) {
4730                                 $hours = round( ( $seconds - $days * 86400 ) / 3600 );
4731                                 if ( $hours == 24 ) {
4732                                         $hours = 0;
4733                                         $days++;
4734                                 }
4735                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4736                                 $s .= ' ';
4737                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4738                         } elseif ( $format['avoid'] === 'avoidseconds' ) {
4739                                 $hours = floor( ( $seconds - $days * 86400 ) / 3600 );
4740                                 $minutes = round( ( $seconds - $days * 86400 - $hours * 3600 ) / 60 );
4741                                 if ( $minutes == 60 ) {
4742                                         $minutes = 0;
4743                                         $hours++;
4744                                 }
4745                                 if ( $hours == 24 ) {
4746                                         $hours = 0;
4747                                         $days++;
4748                                 }
4749                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4750                                 $s .= ' ';
4751                                 $s .= $hoursMsg->params( $this->formatNum( $hours ) )->text();
4752                                 $s .= ' ';
4753                                 $s .= $minutesMsg->params( $this->formatNum( $minutes ) )->text();
4754                         } else {
4755                                 $s = $daysMsg->params( $this->formatNum( $days ) )->text();
4756                                 $s .= ' ';
4757                                 $s .= $this->formatTimePeriod( $seconds - $days * 86400, $format );
4758                         }
4759                 }
4760                 return $s;
4761         }
4762
4763         /**
4764          * Format a bitrate for output, using an appropriate
4765          * unit (bps, kbps, Mbps, Gbps, Tbps, Pbps, Ebps, Zbps or Ybps) according to
4766          *   the magnitude in question.
4767          *
4768          * This use base 1000. For base 1024 use formatSize(), for another base
4769          * see formatComputingNumbers().
4770          *
4771          * @param int $bps
4772          * @return string
4773          */
4774         function formatBitrate( $bps ) {
4775                 return $this->formatComputingNumbers( $bps, 1000, "bitrate-$1bits" );
4776         }
4777
4778         /**
4779          * @param int $size Size of the unit
4780          * @param int $boundary Size boundary (1000, or 1024 in most cases)
4781          * @param string $messageKey Message key to be uesd
4782          * @return string
4783          */
4784         function formatComputingNumbers( $size, $boundary, $messageKey ) {
4785                 if ( $size <= 0 ) {
4786                         return str_replace( '$1', $this->formatNum( $size ),
4787                                 $this->getMessageFromDB( str_replace( '$1', '', $messageKey ) )
4788                         );
4789                 }
4790                 $sizes = [ '', 'kilo', 'mega', 'giga', 'tera', 'peta', 'exa', 'zeta', 'yotta' ];
4791                 $index = 0;
4792
4793                 $maxIndex = count( $sizes ) - 1;
4794                 while ( $size >= $boundary && $index < $maxIndex ) {
4795                         $index++;
4796                         $size /= $boundary;
4797                 }
4798
4799                 // For small sizes no decimal places necessary
4800                 $round = 0;
4801                 if ( $index > 1 ) {
4802                         // For MB and bigger two decimal places are smarter
4803                         $round = 2;
4804                 }
4805                 $msg = str_replace( '$1', $sizes[$index], $messageKey );
4806
4807                 $size = round( $size, $round );
4808                 $text = $this->getMessageFromDB( $msg );
4809                 return str_replace( '$1', $this->formatNum( $size ), $text );
4810         }
4811
4812         /**
4813          * Format a size in bytes for output, using an appropriate
4814          * unit (B, KB, MB, GB, TB, PB, EB, ZB or YB) according to the magnitude in question
4815          *
4816          * This method use base 1024. For base 1000 use formatBitrate(), for
4817          * another base see formatComputingNumbers()
4818          *
4819          * @param int $size Size to format
4820          * @return string Plain text (not HTML)
4821          */
4822         function formatSize( $size ) {
4823                 return $this->formatComputingNumbers( $size, 1024, "size-$1bytes" );
4824         }
4825
4826         /**
4827          * Make a list item, used by various special pages
4828          *
4829          * @param string $page Page link
4830          * @param string $details HTML safe text between brackets
4831          * @param bool $oppositedm Add the direction mark opposite to your
4832          *   language, to display text properly
4833          * @return string HTML escaped
4834          */
4835         function specialList( $page, $details, $oppositedm = true ) {
4836                 if ( !$details ) {
4837                         return $page;
4838                 }
4839
4840                 $dirmark = ( $oppositedm ? $this->getDirMark( true ) : '' ) . $this->getDirMark();
4841                 return $page .
4842                         $dirmark .
4843                         $this->msg( 'word-separator' )->escaped() .
4844                         $this->msg( 'parentheses' )->rawParams( $details )->escaped();
4845         }
4846
4847         /**
4848          * Generate (prev x| next x) (20|50|100...) type links for paging
4849          *
4850          * @param Title $title Title object to link
4851          * @param int $offset
4852          * @param int $limit
4853          * @param array $query Optional URL query parameter string
4854          * @param bool $atend Optional param for specified if this is the last page
4855          * @return string
4856          */
4857         public function viewPrevNext( Title $title, $offset, $limit,
4858                 array $query = [], $atend = false
4859         ) {
4860                 // @todo FIXME: Why on earth this needs one message for the text and another one for tooltip?
4861
4862                 # Make 'previous' link
4863                 $prev = wfMessage( 'prevn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4864                 if ( $offset > 0 ) {
4865                         $plink = $this->numLink( $title, max( $offset - $limit, 0 ), $limit,
4866                                 $query, $prev, 'prevn-title', 'mw-prevlink' );
4867                 } else {
4868                         $plink = htmlspecialchars( $prev );
4869                 }
4870
4871                 # Make 'next' link
4872                 $next = wfMessage( 'nextn' )->inLanguage( $this )->title( $title )->numParams( $limit )->text();
4873                 if ( $atend ) {
4874                         $nlink = htmlspecialchars( $next );
4875                 } else {
4876                         $nlink = $this->numLink( $title, $offset + $limit, $limit,
4877                                 $query, $next, 'nextn-title', 'mw-nextlink' );
4878                 }
4879
4880                 # Make links to set number of items per page
4881                 $numLinks = [];
4882                 foreach ( [ 20, 50, 100, 250, 500 ] as $num ) {
4883                         $numLinks[] = $this->numLink( $title, $offset, $num,
4884                                 $query, $this->formatNum( $num ), 'shown-title', 'mw-numlink' );
4885                 }
4886
4887                 return wfMessage( 'viewprevnext' )->inLanguage( $this )->title( $title
4888                         )->rawParams( $plink, $nlink, $this->pipeList( $numLinks ) )->escaped();
4889         }
4890
4891         /**
4892          * Helper function for viewPrevNext() that generates links
4893          *
4894          * @param Title $title Title object to link
4895          * @param int $offset
4896          * @param int $limit
4897          * @param array $query Extra query parameters
4898          * @param string $link Text to use for the link; will be escaped
4899          * @param string $tooltipMsg Name of the message to use as tooltip
4900          * @param string $class Value of the "class" attribute of the link
4901          * @return string HTML fragment
4902          */
4903         private function numLink( Title $title, $offset, $limit, array $query, $link,
4904                 $tooltipMsg, $class
4905         ) {
4906                 $query = [ 'limit' => $limit, 'offset' => $offset ] + $query;
4907                 $tooltip = wfMessage( $tooltipMsg )->inLanguage( $this )->title( $title )
4908                         ->numParams( $limit )->text();
4909
4910                 return Html::element( 'a', [ 'href' => $title->getLocalURL( $query ),
4911                         'title' => $tooltip, 'class' => $class ], $link );
4912         }
4913
4914         /**
4915          * Get the conversion rule title, if any.
4916          *
4917          * @return string
4918          */
4919         public function getConvRuleTitle() {
4920                 return $this->mConverter->getConvRuleTitle();
4921         }
4922
4923         /**
4924          * Get the compiled plural rules for the language
4925          * @since 1.20
4926          * @return array Associative array with plural form, and plural rule as key-value pairs
4927          */
4928         public function getCompiledPluralRules() {
4929                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'compiledPluralRules' );
4930                 $fallbacks = self::getFallbacksFor( $this->mCode );
4931                 if ( !$pluralRules ) {
4932                         foreach ( $fallbacks as $fallbackCode ) {
4933                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'compiledPluralRules' );
4934                                 if ( $pluralRules ) {
4935                                         break;
4936                                 }
4937                         }
4938                 }
4939                 return $pluralRules;
4940         }
4941
4942         /**
4943          * Get the plural rules for the language
4944          * @since 1.20
4945          * @return array Associative array with plural form number and plural rule as key-value pairs
4946          */
4947         public function getPluralRules() {
4948                 $pluralRules = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRules' );
4949                 $fallbacks = self::getFallbacksFor( $this->mCode );
4950                 if ( !$pluralRules ) {
4951                         foreach ( $fallbacks as $fallbackCode ) {
4952                                 $pluralRules = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRules' );
4953                                 if ( $pluralRules ) {
4954                                         break;
4955                                 }
4956                         }
4957                 }
4958                 return $pluralRules;
4959         }
4960
4961         /**
4962          * Get the plural rule types for the language
4963          * @since 1.22
4964          * @return array Associative array with plural form number and plural rule type as key-value pairs
4965          */
4966         public function getPluralRuleTypes() {
4967                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $this->mCode ), 'pluralRuleTypes' );
4968                 $fallbacks = self::getFallbacksFor( $this->mCode );
4969                 if ( !$pluralRuleTypes ) {
4970                         foreach ( $fallbacks as $fallbackCode ) {
4971                                 $pluralRuleTypes = self::$dataCache->getItem( strtolower( $fallbackCode ), 'pluralRuleTypes' );
4972                                 if ( $pluralRuleTypes ) {
4973                                         break;
4974                                 }
4975                         }
4976                 }
4977                 return $pluralRuleTypes;
4978         }
4979
4980         /**
4981          * Find the index number of the plural rule appropriate for the given number
4982          * @param int $number
4983          * @return int The index number of the plural rule
4984          */
4985         public function getPluralRuleIndexNumber( $number ) {
4986                 $pluralRules = $this->getCompiledPluralRules();
4987                 $form = Evaluator::evaluateCompiled( $number, $pluralRules );
4988                 return $form;
4989         }
4990
4991         /**
4992          * Find the plural rule type appropriate for the given number
4993          * For example, if the language is set to Arabic, getPluralType(5) should
4994          * return 'few'.
4995          * @since 1.22
4996          * @param int $number
4997          * @return string The name of the plural rule type, e.g. one, two, few, many
4998          */
4999         public function getPluralRuleType( $number ) {
5000                 $index = $this->getPluralRuleIndexNumber( $number );
5001                 $pluralRuleTypes = $this->getPluralRuleTypes();
5002                 return $pluralRuleTypes[$index] ?? 'other';
5003         }
5004 }